first

Adafruit_NeoPixel/Adafruit_NeoPixel.cpp

@@ -0,0 +1,2618 @@
+/*!
+ * @file Adafruit_NeoPixel.cpp
+ *
+ * @mainpage Arduino Library for driving Adafruit NeoPixel addressable LEDs,
+ * FLORA RGB Smart Pixels and compatible devicess -- WS2811, WS2812, WS2812B,
+ * SK6812, etc.
+ *
+ * @section intro_sec Introduction
+ *
+ * This is the documentation for Adafruit's NeoPixel library for the
+ * Arduino platform, allowing a broad range of microcontroller boards
+ * (most AVR boards, many ARM devices, ESP8266 and ESP32, among others)
+ * to control Adafruit NeoPixels, FLORA RGB Smart Pixels and compatible
+ * devices -- WS2811, WS2812, WS2812B, SK6812, etc.
+ *
+ * Adafruit invests time and resources providing this open source code,
+ * please support Adafruit and open-source hardware by purchasing products
+ * from Adafruit!
+ *
+ * @section author Author
+ *
+ * Written by Phil "Paint Your Dragon" Burgess for Adafruit Industries,
+ * with contributions by PJRC, Michael Miller and other members of the
+ * open source community.
+ *
+ * @section license License
+ *
+ * This file is part of the Adafruit_NeoPixel library.
+ *
+ * Adafruit_NeoPixel is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public License as
+ * published by the Free Software Foundation, either version 3 of the
+ * License, or (at your option) any later version.
+ *
+ * Adafruit_NeoPixel is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with NeoPixel. If not, see
+ * <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include "Adafruit_NeoPixel.h"
+
+#if defined(TARGET_LPC1768)
+  #include <time.h>
+#endif
+
+#if defined(NRF52) || defined(NRF52_SERIES)
+#include "nrf.h"
+
+// Interrupt is only disabled if there is no PWM device available
+// Note: Adafruit Bluefruit nrf52 does not use this option
+//#define NRF52_DISABLE_INT
+#endif
+
+#if defined(ARDUINO_ARCH_NRF52840)
+#if defined __has_include
+#  if __has_include (<pinDefinitions.h>)
+#    include <pinDefinitions.h>
+#  endif
+#endif
+#endif
+
+/*!
+  @brief   NeoPixel constructor when length, pin and pixel type are known
+           at compile-time.
+  @param   n  Number of NeoPixels in strand.
+  @param   p  Arduino pin number which will drive the NeoPixel data in.
+  @param   t  Pixel type -- add together NEO_* constants defined in
+              Adafruit_NeoPixel.h, for example NEO_GRB+NEO_KHZ800 for
+              NeoPixels expecting an 800 KHz (vs 400 KHz) data stream
+              with color bytes expressed in green, red, blue order per
+              pixel.
+  @return  Adafruit_NeoPixel object. Call the begin() function before use.
+*/
+Adafruit_NeoPixel::Adafruit_NeoPixel(uint16_t n, int16_t p, neoPixelType t) :
+  begun(false), brightness(0), pixels(NULL), endTime(0) {
+  updateType(t);
+  updateLength(n);
+  setPin(p);
+}
+
+/*!
+  @brief   "Empty" NeoPixel constructor when length, pin and/or pixel type
+           are not known at compile-time, and must be initialized later with
+           updateType(), updateLength() and setPin().
+  @return  Adafruit_NeoPixel object. Call the begin() function before use.
+  @note    This function is deprecated, here only for old projects that
+           may still be calling it. New projects should instead use the
+           'new' keyword with the first constructor syntax (length, pin,
+           type).
+*/
+Adafruit_NeoPixel::Adafruit_NeoPixel() :
+#if defined(NEO_KHZ400)
+  is800KHz(true),
+#endif
+  begun(false), numLEDs(0), numBytes(0), pin(-1), brightness(0), pixels(NULL),
+  rOffset(1), gOffset(0), bOffset(2), wOffset(1), endTime(0) {
+}
+
+/*!
+  @brief   Deallocate Adafruit_NeoPixel object, set data pin back to INPUT.
+*/
+Adafruit_NeoPixel::~Adafruit_NeoPixel() {
+  free(pixels);
+  if(pin >= 0) pinMode(pin, INPUT);
+}
+
+/*!
+  @brief   Configure NeoPixel pin for output.
+*/
+void Adafruit_NeoPixel::begin(void) {
+  if(pin >= 0) {
+    pinMode(pin, OUTPUT);
+    digitalWrite(pin, LOW);
+  }
+  begun = true;
+}
+
+/*!
+  @brief   Change the length of a previously-declared Adafruit_NeoPixel
+           strip object. Old data is deallocated and new data is cleared.
+           Pin number and pixel format are unchanged.
+  @param   n  New length of strip, in pixels.
+  @note    This function is deprecated, here only for old projects that
+           may still be calling it. New projects should instead use the
+           'new' keyword with the first constructor syntax (length, pin,
+           type).
+*/
+void Adafruit_NeoPixel::updateLength(uint16_t n) {
+  free(pixels); // Free existing data (if any)
+
+  // Allocate new data -- note: ALL PIXELS ARE CLEARED
+  numBytes = n * ((wOffset == rOffset) ? 3 : 4);
+  if((pixels = (uint8_t *)malloc(numBytes))) {
+    memset(pixels, 0, numBytes);
+    numLEDs = n;
+  } else {
+    numLEDs = numBytes = 0;
+  }
+}
+
+/*!
+  @brief   Change the pixel format of a previously-declared
+           Adafruit_NeoPixel strip object. If format changes from one of
+           the RGB variants to an RGBW variant (or RGBW to RGB), the old
+           data will be deallocated and new data is cleared. Otherwise,
+           the old data will remain in RAM and is not reordered to the
+           new format, so it's advisable to follow up with clear().
+  @param   t  Pixel type -- add together NEO_* constants defined in
+              Adafruit_NeoPixel.h, for example NEO_GRB+NEO_KHZ800 for
+              NeoPixels expecting an 800 KHz (vs 400 KHz) data stream
+              with color bytes expressed in green, red, blue order per
+              pixel.
+  @note    This function is deprecated, here only for old projects that
+           may still be calling it. New projects should instead use the
+           'new' keyword with the first constructor syntax
+           (length, pin, type).
+*/
+void Adafruit_NeoPixel::updateType(neoPixelType t) {
+  bool oldThreeBytesPerPixel = (wOffset == rOffset); // false if RGBW
+
+  wOffset = (t >> 6) & 0b11; // See notes in header file
+  rOffset = (t >> 4) & 0b11; // regarding R/G/B/W offsets
+  gOffset = (t >> 2) & 0b11;
+  bOffset =  t       & 0b11;
+#if defined(NEO_KHZ400)
+  is800KHz = (t < 256);      // 400 KHz flag is 1<<8
+#endif
+
+  // If bytes-per-pixel has changed (and pixel data was previously
+  // allocated), re-allocate to new size. Will clear any data.
+  if(pixels) {
+    bool newThreeBytesPerPixel = (wOffset == rOffset);
+    if(newThreeBytesPerPixel != oldThreeBytesPerPixel) updateLength(numLEDs);
+  }
+}
+
+#if defined(ARDUINO_ARCH_RP2040)
+extern "C" void rp2040Show(
+  uint16_t pin, uint8_t *pixels, uint32_t numBytes, uint8_t type);
+#endif
+
+#if defined(ESP8266)
+// ESP8266 show() is external to enforce ICACHE_RAM_ATTR execution
+extern "C" IRAM_ATTR void espShow(
+  uint16_t pin, uint8_t *pixels, uint32_t numBytes, uint8_t type);
+#elif defined(ESP32)
+extern "C" void espShow(
+  uint16_t pin, uint8_t *pixels, uint32_t numBytes, uint8_t type);
+#endif // ESP8266
+
+#if defined(K210) 
+#define KENDRYTE_K210 1
+#endif
+
+#if defined(KENDRYTE_K210)
+extern "C" void  k210Show(
+    uint8_t pin, uint8_t *pixels, uint32_t numBytes, boolean is800KHz);
+#endif //KENDRYTE_K210
+/*!
+  @brief   Transmit pixel data in RAM to NeoPixels.
+  @note    On most architectures, interrupts are temporarily disabled in
+           order to achieve the correct NeoPixel signal timing. This means
+           that the Arduino millis() and micros() functions, which require
+           interrupts, will lose small intervals of time whenever this
+           function is called (about 30 microseconds per RGB pixel, 40 for
+           RGBW pixels). There's no easy fix for this, but a few
+           specialized alternative or companion libraries exist that use
+           very device-specific peripherals to work around it.
+*/
+void Adafruit_NeoPixel::show(void) {
+
+  if(!pixels) return;
+
+  // Data latch = 300+ microsecond pause in the output stream. Rather than
+  // put a delay at the end of the function, the ending time is noted and
+  // the function will simply hold off (if needed) on issuing the
+  // subsequent round of data until the latch time has elapsed. This
+  // allows the mainline code to start generating the next frame of data
+  // rather than stalling for the latch.
+  while(!canShow());
+  // endTime is a private member (rather than global var) so that multiple
+  // instances on different pins can be quickly issued in succession (each
+  // instance doesn't delay the next).
+
+  // In order to make this code runtime-configurable to work with any pin,
+  // SBI/CBI instructions are eschewed in favor of full PORT writes via the
+  // OUT or ST instructions. It relies on two facts: that peripheral
+  // functions (such as PWM) take precedence on output pins, so our PORT-
+  // wide writes won't interfere, and that interrupts are globally disabled
+  // while data is being issued to the LEDs, so no other code will be
+  // accessing the PORT. The code takes an initial 'snapshot' of the PORT
+  // state, computes 'pin high' and 'pin low' values, and writes these back
+  // to the PORT register as needed.
+
+  // NRF52 may use PWM + DMA (if available), may not need to disable interrupt
+#if !( defined(NRF52) || defined(NRF52_SERIES) )
+  noInterrupts(); // Need 100% focus on instruction timing
+#endif
+
+#if defined(__AVR__)
+// AVR MCUs -- ATmega & ATtiny (no XMEGA) ---------------------------------
+
+  volatile uint16_t
+    i   = numBytes; // Loop counter
+  volatile uint8_t
+   *ptr = pixels,   // Pointer to next byte
+    b   = *ptr++,   // Current byte value
+    hi,             // PORT w/output bit set high
+    lo;             // PORT w/output bit set low
+
+  // Hand-tuned assembly code issues data to the LED drivers at a specific
+  // rate. There's separate code for different CPU speeds (8, 12, 16 MHz)
+  // for both the WS2811 (400 KHz) and WS2812 (800 KHz) drivers. The
+  // datastream timing for the LED drivers allows a little wiggle room each
+  // way (listed in the datasheets), so the conditions for compiling each
+  // case are set up for a range of frequencies rather than just the exact
+  // 8, 12 or 16 MHz values, permitting use with some close-but-not-spot-on
+  // devices (e.g. 16.5 MHz DigiSpark). The ranges were arrived at based
+  // on the datasheet figures and have not been extensively tested outside
+  // the canonical 8/12/16 MHz speeds; there's no guarantee these will work
+  // close to the extremes (or possibly they could be pushed further).
+  // Keep in mind only one CPU speed case actually gets compiled; the
+  // resulting program isn't as massive as it might look from source here.
+
+// 8 MHz(ish) AVR ---------------------------------------------------------
+#if (F_CPU >= 7400000UL) && (F_CPU <= 9500000UL)
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+
+    volatile uint8_t n1, n2 = 0;  // First, next bits out
+
+    // Squeezing an 800 KHz stream out of an 8 MHz chip requires code
+    // specific to each PORT register.
+
+    // 10 instruction clocks per bit: HHxxxxxLLL
+    // OUT instructions:              ^ ^    ^   (T=0,2,7)
+
+    // PORTD OUTPUT ----------------------------------------------------
+
+#if defined(PORTD)
+ #if defined(PORTB) || defined(PORTC) || defined(PORTF)
+    if(port == &PORTD) {
+ #endif // defined(PORTB/C/F)
+
+      hi = PORTD |  pinMask;
+      lo = PORTD & ~pinMask;
+      n1 = lo;
+      if(b & 0x80) n1 = hi;
+
+      // Dirty trick: RJMPs proceeding to the next instruction are used
+      // to delay two clock cycles in one instruction word (rather than
+      // using two NOPs). This was necessary in order to squeeze the
+      // loop down to exactly 64 words -- the maximum possible for a
+      // relative branch.
+
+      asm volatile(
+       "headD:"                   "\n\t" // Clk  Pseudocode
+        // Bit 7:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n2]   , %[lo]"    "\n\t" // 1    n2   = lo
+        "out  %[port] , %[n1]"    "\n\t" // 1    PORT = n1
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 6"        "\n\t" // 1-2  if(b & 0x40)
+         "mov %[n2]   , %[hi]"    "\n\t" // 0-1   n2 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 6:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n1]   , %[lo]"    "\n\t" // 1    n1   = lo
+        "out  %[port] , %[n2]"    "\n\t" // 1    PORT = n2
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 5"        "\n\t" // 1-2  if(b & 0x20)
+         "mov %[n1]   , %[hi]"    "\n\t" // 0-1   n1 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 5:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n2]   , %[lo]"    "\n\t" // 1    n2   = lo
+        "out  %[port] , %[n1]"    "\n\t" // 1    PORT = n1
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 4"        "\n\t" // 1-2  if(b & 0x10)
+         "mov %[n2]   , %[hi]"    "\n\t" // 0-1   n2 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 4:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n1]   , %[lo]"    "\n\t" // 1    n1   = lo
+        "out  %[port] , %[n2]"    "\n\t" // 1    PORT = n2
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 3"        "\n\t" // 1-2  if(b & 0x08)
+         "mov %[n1]   , %[hi]"    "\n\t" // 0-1   n1 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 3:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n2]   , %[lo]"    "\n\t" // 1    n2   = lo
+        "out  %[port] , %[n1]"    "\n\t" // 1    PORT = n1
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 2"        "\n\t" // 1-2  if(b & 0x04)
+         "mov %[n2]   , %[hi]"    "\n\t" // 0-1   n2 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 2:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n1]   , %[lo]"    "\n\t" // 1    n1   = lo
+        "out  %[port] , %[n2]"    "\n\t" // 1    PORT = n2
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 1"        "\n\t" // 1-2  if(b & 0x02)
+         "mov %[n1]   , %[hi]"    "\n\t" // 0-1   n1 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        // Bit 1:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n2]   , %[lo]"    "\n\t" // 1    n2   = lo
+        "out  %[port] , %[n1]"    "\n\t" // 1    PORT = n1
+        "rjmp .+0"                "\n\t" // 2    nop nop
+        "sbrc %[byte] , 0"        "\n\t" // 1-2  if(b & 0x01)
+         "mov %[n2]   , %[hi]"    "\n\t" // 0-1   n2 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "sbiw %[count], 1"        "\n\t" // 2    i-- (don't act on Z flag yet)
+        // Bit 0:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi
+        "mov  %[n1]   , %[lo]"    "\n\t" // 1    n1   = lo
+        "out  %[port] , %[n2]"    "\n\t" // 1    PORT = n2
+        "ld   %[byte] , %a[ptr]+" "\n\t" // 2    b = *ptr++
+        "sbrc %[byte] , 7"        "\n\t" // 1-2  if(b & 0x80)
+         "mov %[n1]   , %[hi]"    "\n\t" // 0-1   n1 = hi
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo
+        "brne headD"              "\n"   // 2    while(i) (Z flag set above)
+      : [byte]  "+r" (b),
+        [n1]    "+r" (n1),
+        [n2]    "+r" (n2),
+        [count] "+w" (i)
+      : [port]   "I" (_SFR_IO_ADDR(PORTD)),
+        [ptr]    "e" (ptr),
+        [hi]     "r" (hi),
+        [lo]     "r" (lo));
+
+ #if defined(PORTB) || defined(PORTC) || defined(PORTF)
+    } else // other PORT(s)
+ #endif // defined(PORTB/C/F)
+#endif // defined(PORTD)
+
+    // PORTB OUTPUT ----------------------------------------------------
+
+#if defined(PORTB)
+ #if defined(PORTD) || defined(PORTC) || defined(PORTF)
+    if(port == &PORTB) {
+ #endif // defined(PORTD/C/F)
+
+      // Same as above, just switched to PORTB and stripped of comments.
+      hi = PORTB |  pinMask;
+      lo = PORTB & ~pinMask;
+      n1 = lo;
+      if(b & 0x80) n1 = hi;
+
+      asm volatile(
+       "headB:"                   "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 6"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 5"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 4"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 3"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 2"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 1"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 0"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "brne headB"              "\n"
+      : [byte] "+r" (b), [n1] "+r" (n1), [n2] "+r" (n2), [count] "+w" (i)
+      : [port] "I" (_SFR_IO_ADDR(PORTB)), [ptr] "e" (ptr), [hi] "r" (hi),
+        [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTC) || defined(PORTF)
+    }
+ #endif
+ #if defined(PORTC) || defined(PORTF)
+    else
+ #endif // defined(PORTC/F)
+#endif // defined(PORTB)
+
+    // PORTC OUTPUT ----------------------------------------------------
+
+#if defined(PORTC)
+ #if defined(PORTD) || defined(PORTB) || defined(PORTF)
+    if(port == &PORTC) {
+ #endif // defined(PORTD/B/F)
+
+      // Same as above, just switched to PORTC and stripped of comments.
+      hi = PORTC |  pinMask;
+      lo = PORTC & ~pinMask;
+      n1 = lo;
+      if(b & 0x80) n1 = hi;
+
+      asm volatile(
+       "headC:"                   "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 6"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 5"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 4"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 3"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 2"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 1"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 0"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "brne headC"              "\n"
+      : [byte] "+r" (b), [n1] "+r" (n1), [n2] "+r" (n2), [count] "+w" (i)
+      : [port] "I" (_SFR_IO_ADDR(PORTC)), [ptr] "e" (ptr), [hi] "r" (hi),
+        [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTB) || defined(PORTF)
+    }
+ #endif // defined(PORTD/B/F)
+ #if defined(PORTF)
+    else
+ #endif
+#endif // defined(PORTC)
+
+    // PORTF OUTPUT ----------------------------------------------------
+
+#if defined(PORTF)
+ #if defined(PORTD) || defined(PORTB) || defined(PORTC)
+    if(port == &PORTF) {
+ #endif // defined(PORTD/B/C)
+
+      hi = PORTF |  pinMask;
+      lo = PORTF & ~pinMask;
+      n1 = lo;
+      if(b & 0x80) n1 = hi;
+
+      asm volatile(
+       "headF:"                   "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 6"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 5"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 4"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 3"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 2"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 1"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n2]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n1]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "sbrc %[byte] , 0"        "\n\t"
+         "mov %[n2]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "mov  %[n1]   , %[lo]"    "\n\t"
+        "out  %[port] , %[n2]"    "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[n1]   , %[hi]"    "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "brne headF"              "\n"
+      : [byte] "+r" (b), [n1] "+r" (n1), [n2] "+r" (n2), [count] "+w" (i)
+      : [port] "I" (_SFR_IO_ADDR(PORTF)), [ptr] "e" (ptr), [hi] "r" (hi),
+        [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTB) || defined(PORTC)
+    }
+ #endif // defined(PORTD/B/C)
+#endif // defined(PORTF)
+
+#if defined(NEO_KHZ400)
+  } else { // end 800 KHz, do 400 KHz
+
+    // Timing is more relaxed; unrolling the inner loop for each bit is
+    // not necessary. Still using the peculiar RJMPs as 2X NOPs, not out
+    // of need but just to trim the code size down a little.
+    // This 400-KHz-datastream-on-8-MHz-CPU code is not quite identical
+    // to the 800-on-16 code later -- the hi/lo timing between WS2811 and
+    // WS2812 is not simply a 2:1 scale!
+
+    // 20 inst. clocks per bit: HHHHxxxxxxLLLLLLLLLL
+    // ST instructions:         ^   ^     ^          (T=0,4,10)
+
+    volatile uint8_t next, bit;
+
+    hi   = *port |  pinMask;
+    lo   = *port & ~pinMask;
+    next = lo;
+    bit  = 8;
+
+    asm volatile(
+     "head20:"                  "\n\t" // Clk  Pseudocode    (T =  0)
+      "st   %a[port], %[hi]"    "\n\t" // 2    PORT = hi     (T =  2)
+      "sbrc %[byte] , 7"        "\n\t" // 1-2  if(b & 128)
+       "mov  %[next], %[hi]"    "\n\t" // 0-1   next = hi    (T =  4)
+      "st   %a[port], %[next]"  "\n\t" // 2    PORT = next   (T =  6)
+      "mov  %[next] , %[lo]"    "\n\t" // 1    next = lo     (T =  7)
+      "dec  %[bit]"             "\n\t" // 1    bit--         (T =  8)
+      "breq nextbyte20"         "\n\t" // 1-2  if(bit == 0)
+      "rol  %[byte]"            "\n\t" // 1    b <<= 1       (T = 10)
+      "st   %a[port], %[lo]"    "\n\t" // 2    PORT = lo     (T = 12)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 14)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 16)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 18)
+      "rjmp head20"             "\n\t" // 2    -> head20 (next bit out)
+     "nextbyte20:"              "\n\t" //                    (T = 10)
+      "st   %a[port], %[lo]"    "\n\t" // 2    PORT = lo     (T = 12)
+      "nop"                     "\n\t" // 1    nop           (T = 13)
+      "ldi  %[bit]  , 8"        "\n\t" // 1    bit = 8       (T = 14)
+      "ld   %[byte] , %a[ptr]+" "\n\t" // 2    b = *ptr++    (T = 16)
+      "sbiw %[count], 1"        "\n\t" // 2    i--           (T = 18)
+      "brne head20"             "\n"   // 2    if(i != 0) -> (next byte)
+      : [port]  "+e" (port),
+        [byte]  "+r" (b),
+        [bit]   "+r" (bit),
+        [next]  "+r" (next),
+        [count] "+w" (i)
+      : [hi]    "r" (hi),
+        [lo]    "r" (lo),
+        [ptr]   "e" (ptr));
+  }
+#endif // NEO_KHZ400
+
+// 12 MHz(ish) AVR --------------------------------------------------------
+#elif (F_CPU >= 11100000UL) && (F_CPU <= 14300000UL)
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+
+    // In the 12 MHz case, an optimized 800 KHz datastream (no dead time
+    // between bytes) requires a PORT-specific loop similar to the 8 MHz
+    // code (but a little more relaxed in this case).
+
+    // 15 instruction clocks per bit: HHHHxxxxxxLLLLL
+    // OUT instructions:              ^   ^     ^     (T=0,4,10)
+
+    volatile uint8_t next;
+
+    // PORTD OUTPUT ----------------------------------------------------
+
+#if defined(PORTD)
+ #if defined(PORTB) || defined(PORTC) || defined(PORTF)
+    if(port == &PORTD) {
+ #endif // defined(PORTB/C/F)
+
+      hi   = PORTD |  pinMask;
+      lo   = PORTD & ~pinMask;
+      next = lo;
+      if(b & 0x80) next = hi;
+
+      // Don't "optimize" the OUT calls into the bitTime subroutine;
+      // we're exploiting the RCALL and RET as 3- and 4-cycle NOPs!
+      asm volatile(
+       "headD:"                   "\n\t" //        (T =  0)
+        "out   %[port], %[hi]"    "\n\t" //        (T =  1)
+        "rcall bitTimeD"          "\n\t" // Bit 7  (T = 15)
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 6
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 5
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 4
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 3
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 2
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeD"          "\n\t" // Bit 1
+        // Bit 0:
+        "out  %[port] , %[hi]"    "\n\t" // 1    PORT = hi    (T =  1)
+        "rjmp .+0"                "\n\t" // 2    nop nop      (T =  3)
+        "ld   %[byte] , %a[ptr]+" "\n\t" // 2    b = *ptr++   (T =  5)
+        "out  %[port] , %[next]"  "\n\t" // 1    PORT = next  (T =  6)
+        "mov  %[next] , %[lo]"    "\n\t" // 1    next = lo    (T =  7)
+        "sbrc %[byte] , 7"        "\n\t" // 1-2  if(b & 0x80) (T =  8)
+         "mov %[next] , %[hi]"    "\n\t" // 0-1    next = hi  (T =  9)
+        "nop"                     "\n\t" // 1                 (T = 10)
+        "out  %[port] , %[lo]"    "\n\t" // 1    PORT = lo    (T = 11)
+        "sbiw %[count], 1"        "\n\t" // 2    i--          (T = 13)
+        "brne headD"              "\n\t" // 2    if(i != 0) -> (next byte)
+         "rjmp doneD"             "\n\t"
+        "bitTimeD:"               "\n\t" //      nop nop nop     (T =  4)
+         "out  %[port], %[next]"  "\n\t" // 1    PORT = next     (T =  5)
+         "mov  %[next], %[lo]"    "\n\t" // 1    next = lo       (T =  6)
+         "rol  %[byte]"           "\n\t" // 1    b <<= 1         (T =  7)
+         "sbrc %[byte], 7"        "\n\t" // 1-2  if(b & 0x80)    (T =  8)
+          "mov %[next], %[hi]"    "\n\t" // 0-1   next = hi      (T =  9)
+         "nop"                    "\n\t" // 1                    (T = 10)
+         "out  %[port], %[lo]"    "\n\t" // 1    PORT = lo       (T = 11)
+         "ret"                    "\n\t" // 4    nop nop nop nop (T = 15)
+         "doneD:"                 "\n"
+        : [byte]  "+r" (b),
+          [next]  "+r" (next),
+          [count] "+w" (i)
+        : [port]   "I" (_SFR_IO_ADDR(PORTD)),
+          [ptr]    "e" (ptr),
+          [hi]     "r" (hi),
+          [lo]     "r" (lo));
+
+ #if defined(PORTB) || defined(PORTC) || defined(PORTF)
+    } else // other PORT(s)
+ #endif // defined(PORTB/C/F)
+#endif // defined(PORTD)
+
+    // PORTB OUTPUT ----------------------------------------------------
+
+#if defined(PORTB)
+ #if defined(PORTD) || defined(PORTC) || defined(PORTF)
+    if(port == &PORTB) {
+ #endif // defined(PORTD/C/F)
+
+      hi   = PORTB |  pinMask;
+      lo   = PORTB & ~pinMask;
+      next = lo;
+      if(b & 0x80) next = hi;
+
+      // Same as above, just set for PORTB & stripped of comments
+      asm volatile(
+       "headB:"                   "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeB"          "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "out  %[port] , %[next]"  "\n\t"
+        "mov  %[next] , %[lo]"    "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[next] , %[hi]"    "\n\t"
+        "nop"                     "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "brne headB"              "\n\t"
+         "rjmp doneB"             "\n\t"
+        "bitTimeB:"               "\n\t"
+         "out  %[port], %[next]"  "\n\t"
+         "mov  %[next], %[lo]"    "\n\t"
+         "rol  %[byte]"           "\n\t"
+         "sbrc %[byte], 7"        "\n\t"
+          "mov %[next], %[hi]"    "\n\t"
+         "nop"                    "\n\t"
+         "out  %[port], %[lo]"    "\n\t"
+         "ret"                    "\n\t"
+         "doneB:"                 "\n"
+        : [byte] "+r" (b), [next] "+r" (next), [count] "+w" (i)
+        : [port] "I" (_SFR_IO_ADDR(PORTB)), [ptr] "e" (ptr), [hi] "r" (hi),
+          [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTC) || defined(PORTF)
+    }
+ #endif
+ #if defined(PORTC) || defined(PORTF)
+    else
+ #endif // defined(PORTC/F)
+#endif // defined(PORTB)
+
+    // PORTC OUTPUT ----------------------------------------------------
+
+#if defined(PORTC)
+ #if defined(PORTD) || defined(PORTB) || defined(PORTF)
+    if(port == &PORTC) {
+ #endif // defined(PORTD/B/F)
+
+      hi   = PORTC |  pinMask;
+      lo   = PORTC & ~pinMask;
+      next = lo;
+      if(b & 0x80) next = hi;
+
+      // Same as above, just set for PORTC & stripped of comments
+      asm volatile(
+       "headC:"                   "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "out  %[port] , %[next]"  "\n\t"
+        "mov  %[next] , %[lo]"    "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[next] , %[hi]"    "\n\t"
+        "nop"                     "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "brne headC"              "\n\t"
+         "rjmp doneC"             "\n\t"
+        "bitTimeC:"               "\n\t"
+         "out  %[port], %[next]"  "\n\t"
+         "mov  %[next], %[lo]"    "\n\t"
+         "rol  %[byte]"           "\n\t"
+         "sbrc %[byte], 7"        "\n\t"
+          "mov %[next], %[hi]"    "\n\t"
+         "nop"                    "\n\t"
+         "out  %[port], %[lo]"    "\n\t"
+         "ret"                    "\n\t"
+         "doneC:"                 "\n"
+        : [byte] "+r" (b), [next] "+r" (next), [count] "+w" (i)
+        : [port] "I" (_SFR_IO_ADDR(PORTC)), [ptr] "e" (ptr), [hi] "r" (hi),
+          [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTB) || defined(PORTF)
+    }
+ #endif // defined(PORTD/B/F)
+ #if defined(PORTF)
+    else
+ #endif
+#endif // defined(PORTC)
+
+    // PORTF OUTPUT ----------------------------------------------------
+
+#if defined(PORTF)
+ #if defined(PORTD) || defined(PORTB) || defined(PORTC)
+    if(port == &PORTF) {
+ #endif // defined(PORTD/B/C)
+
+      hi   = PORTF |  pinMask;
+      lo   = PORTF & ~pinMask;
+      next = lo;
+      if(b & 0x80) next = hi;
+
+      // Same as above, just set for PORTF & stripped of comments
+      asm volatile(
+       "headF:"                   "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out   %[port], %[hi]"    "\n\t"
+        "rcall bitTimeC"          "\n\t"
+        "out  %[port] , %[hi]"    "\n\t"
+        "rjmp .+0"                "\n\t"
+        "ld   %[byte] , %a[ptr]+" "\n\t"
+        "out  %[port] , %[next]"  "\n\t"
+        "mov  %[next] , %[lo]"    "\n\t"
+        "sbrc %[byte] , 7"        "\n\t"
+         "mov %[next] , %[hi]"    "\n\t"
+        "nop"                     "\n\t"
+        "out  %[port] , %[lo]"    "\n\t"
+        "sbiw %[count], 1"        "\n\t"
+        "brne headF"              "\n\t"
+         "rjmp doneC"             "\n\t"
+        "bitTimeC:"               "\n\t"
+         "out  %[port], %[next]"  "\n\t"
+         "mov  %[next], %[lo]"    "\n\t"
+         "rol  %[byte]"           "\n\t"
+         "sbrc %[byte], 7"        "\n\t"
+          "mov %[next], %[hi]"    "\n\t"
+         "nop"                    "\n\t"
+         "out  %[port], %[lo]"    "\n\t"
+         "ret"                    "\n\t"
+         "doneC:"                 "\n"
+        : [byte] "+r" (b), [next] "+r" (next), [count] "+w" (i)
+        : [port] "I" (_SFR_IO_ADDR(PORTF)), [ptr] "e" (ptr), [hi] "r" (hi),
+          [lo] "r" (lo));
+
+ #if defined(PORTD) || defined(PORTB) || defined(PORTC)
+    }
+ #endif // defined(PORTD/B/C)
+#endif // defined(PORTF)
+
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz
+
+    // 30 instruction clocks per bit: HHHHHHxxxxxxxxxLLLLLLLLLLLLLLL
+    // ST instructions:               ^     ^        ^    (T=0,6,15)
+
+    volatile uint8_t next, bit;
+
+    hi   = *port |  pinMask;
+    lo   = *port & ~pinMask;
+    next = lo;
+    bit  = 8;
+
+    asm volatile(
+     "head30:"                  "\n\t" // Clk  Pseudocode    (T =  0)
+      "st   %a[port], %[hi]"    "\n\t" // 2    PORT = hi     (T =  2)
+      "sbrc %[byte] , 7"        "\n\t" // 1-2  if(b & 128)
+       "mov  %[next], %[hi]"    "\n\t" // 0-1   next = hi    (T =  4)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T =  6)
+      "st   %a[port], %[next]"  "\n\t" // 2    PORT = next   (T =  8)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 10)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 12)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 14)
+      "nop"                     "\n\t" // 1    nop           (T = 15)
+      "st   %a[port], %[lo]"    "\n\t" // 2    PORT = lo     (T = 17)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 19)
+      "dec  %[bit]"             "\n\t" // 1    bit--         (T = 20)
+      "breq nextbyte30"         "\n\t" // 1-2  if(bit == 0)
+      "rol  %[byte]"            "\n\t" // 1    b <<= 1       (T = 22)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 24)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 26)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 28)
+      "rjmp head30"             "\n\t" // 2    -> head30 (next bit out)
+     "nextbyte30:"              "\n\t" //                    (T = 22)
+      "nop"                     "\n\t" // 1    nop           (T = 23)
+      "ldi  %[bit]  , 8"        "\n\t" // 1    bit = 8       (T = 24)
+      "ld   %[byte] , %a[ptr]+" "\n\t" // 2    b = *ptr++    (T = 26)
+      "sbiw %[count], 1"        "\n\t" // 2    i--           (T = 28)
+      "brne head30"             "\n"   // 1-2  if(i != 0) -> (next byte)
+      : [port]  "+e" (port),
+        [byte]  "+r" (b),
+        [bit]   "+r" (bit),
+        [next]  "+r" (next),
+        [count] "+w" (i)
+      : [hi]     "r" (hi),
+        [lo]     "r" (lo),
+        [ptr]    "e" (ptr));
+  }
+#endif // NEO_KHZ400
+
+// 16 MHz(ish) AVR --------------------------------------------------------
+#elif (F_CPU >= 15400000UL) && (F_CPU <= 19000000L)
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+
+    // WS2811 and WS2812 have different hi/lo duty cycles; this is
+    // similar but NOT an exact copy of the prior 400-on-8 code.
+
+    // 20 inst. clocks per bit: HHHHHxxxxxxxxLLLLLLL
+    // ST instructions:         ^   ^        ^       (T=0,5,13)
+
+    volatile uint8_t next, bit;
+
+    hi   = *port |  pinMask;
+    lo   = *port & ~pinMask;
+    next = lo;
+    bit  = 8;
+
+    asm volatile(
+     "head20:"                   "\n\t" // Clk  Pseudocode    (T =  0)
+      "st   %a[port],  %[hi]"    "\n\t" // 2    PORT = hi     (T =  2)
+      "sbrc %[byte],  7"         "\n\t" // 1-2  if(b & 128)
+       "mov  %[next], %[hi]"     "\n\t" // 0-1   next = hi    (T =  4)
+      "dec  %[bit]"              "\n\t" // 1    bit--         (T =  5)
+      "st   %a[port],  %[next]"  "\n\t" // 2    PORT = next   (T =  7)
+      "mov  %[next] ,  %[lo]"    "\n\t" // 1    next = lo     (T =  8)
+      "breq nextbyte20"          "\n\t" // 1-2  if(bit == 0) (from dec above)
+      "rol  %[byte]"             "\n\t" // 1    b <<= 1       (T = 10)
+      "rjmp .+0"                 "\n\t" // 2    nop nop       (T = 12)
+      "nop"                      "\n\t" // 1    nop           (T = 13)
+      "st   %a[port],  %[lo]"    "\n\t" // 2    PORT = lo     (T = 15)
+      "nop"                      "\n\t" // 1    nop           (T = 16)
+      "rjmp .+0"                 "\n\t" // 2    nop nop       (T = 18)
+      "rjmp head20"              "\n\t" // 2    -> head20 (next bit out)
+     "nextbyte20:"               "\n\t" //                    (T = 10)
+      "ldi  %[bit]  ,  8"        "\n\t" // 1    bit = 8       (T = 11)
+      "ld   %[byte] ,  %a[ptr]+" "\n\t" // 2    b = *ptr++    (T = 13)
+      "st   %a[port], %[lo]"     "\n\t" // 2    PORT = lo     (T = 15)
+      "nop"                      "\n\t" // 1    nop           (T = 16)
+      "sbiw %[count], 1"         "\n\t" // 2    i--           (T = 18)
+       "brne head20"             "\n"   // 2    if(i != 0) -> (next byte)
+      : [port]  "+e" (port),
+        [byte]  "+r" (b),
+        [bit]   "+r" (bit),
+        [next]  "+r" (next),
+        [count] "+w" (i)
+      : [ptr]    "e" (ptr),
+        [hi]     "r" (hi),
+        [lo]     "r" (lo));
+
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz
+
+    // The 400 KHz clock on 16 MHz MCU is the most 'relaxed' version.
+
+    // 40 inst. clocks per bit: HHHHHHHHxxxxxxxxxxxxLLLLLLLLLLLLLLLLLLLL
+    // ST instructions:         ^       ^           ^         (T=0,8,20)
+
+    volatile uint8_t next, bit;
+
+    hi   = *port |  pinMask;
+    lo   = *port & ~pinMask;
+    next = lo;
+    bit  = 8;
+
+    asm volatile(
+     "head40:"                  "\n\t" // Clk  Pseudocode    (T =  0)
+      "st   %a[port], %[hi]"    "\n\t" // 2    PORT = hi     (T =  2)
+      "sbrc %[byte] , 7"        "\n\t" // 1-2  if(b & 128)
+       "mov  %[next] , %[hi]"   "\n\t" // 0-1   next = hi    (T =  4)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T =  6)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T =  8)
+      "st   %a[port], %[next]"  "\n\t" // 2    PORT = next   (T = 10)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 12)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 14)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 16)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 18)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 20)
+      "st   %a[port], %[lo]"    "\n\t" // 2    PORT = lo     (T = 22)
+      "nop"                     "\n\t" // 1    nop           (T = 23)
+      "mov  %[next] , %[lo]"    "\n\t" // 1    next = lo     (T = 24)
+      "dec  %[bit]"             "\n\t" // 1    bit--         (T = 25)
+      "breq nextbyte40"         "\n\t" // 1-2  if(bit == 0)
+      "rol  %[byte]"            "\n\t" // 1    b <<= 1       (T = 27)
+      "nop"                     "\n\t" // 1    nop           (T = 28)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 30)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 32)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 34)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 36)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 38)
+      "rjmp head40"             "\n\t" // 2    -> head40 (next bit out)
+     "nextbyte40:"              "\n\t" //                    (T = 27)
+      "ldi  %[bit]  , 8"        "\n\t" // 1    bit = 8       (T = 28)
+      "ld   %[byte] , %a[ptr]+" "\n\t" // 2    b = *ptr++    (T = 30)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 32)
+      "st   %a[port], %[lo]"    "\n\t" // 2    PORT = lo     (T = 34)
+      "rjmp .+0"                "\n\t" // 2    nop nop       (T = 36)
+      "sbiw %[count], 1"        "\n\t" // 2    i--           (T = 38)
+      "brne head40"             "\n"   // 1-2  if(i != 0) -> (next byte)
+      : [port]  "+e" (port),
+        [byte]  "+r" (b),
+        [bit]   "+r" (bit),
+        [next]  "+r" (next),
+        [count] "+w" (i)
+      : [ptr]    "e" (ptr),
+        [hi]     "r" (hi),
+        [lo]     "r" (lo));
+  }
+#endif // NEO_KHZ400
+
+#else
+ #error "CPU SPEED NOT SUPPORTED"
+#endif // end F_CPU ifdefs on __AVR__
+
+// END AVR ----------------------------------------------------------------
+
+
+#elif defined(__arm__)
+
+// ARM MCUs -- Teensy 3.0, 3.1, LC, Arduino Due, RP2040 -------------------
+
+#if defined(ARDUINO_ARCH_RP2040)
+  // Use PIO
+  rp2040Show(pin, pixels, numBytes, is800KHz);
+
+#elif defined(TEENSYDUINO) && defined(KINETISK) // Teensy 3.0, 3.1, 3.2, 3.5, 3.6
+#define CYCLES_800_T0H  (F_CPU / 4000000)
+#define CYCLES_800_T1H  (F_CPU / 1250000)
+#define CYCLES_800      (F_CPU /  800000)
+#define CYCLES_400_T0H  (F_CPU / 2000000)
+#define CYCLES_400_T1H  (F_CPU /  833333)
+#define CYCLES_400      (F_CPU /  400000)
+
+  uint8_t          *p   = pixels,
+                   *end = p + numBytes, pix, mask;
+  volatile uint8_t *set = portSetRegister(pin),
+                   *clr = portClearRegister(pin);
+  uint32_t          cyc;
+
+  ARM_DEMCR    |= ARM_DEMCR_TRCENA;
+  ARM_DWT_CTRL |= ARM_DWT_CTRL_CYCCNTENA;
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    cyc = ARM_DWT_CYCCNT + CYCLES_800;
+    while(p < end) {
+      pix = *p++;
+      for(mask = 0x80; mask; mask >>= 1) {
+        while(ARM_DWT_CYCCNT - cyc < CYCLES_800);
+        cyc  = ARM_DWT_CYCCNT;
+        *set = 1;
+        if(pix & mask) {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_800_T1H);
+        } else {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_800_T0H);
+        }
+        *clr = 1;
+      }
+    }
+    while(ARM_DWT_CYCCNT - cyc < CYCLES_800);
+#if defined(NEO_KHZ400)
+  } else { // 400 kHz bitstream
+    cyc = ARM_DWT_CYCCNT + CYCLES_400;
+    while(p < end) {
+      pix = *p++;
+      for(mask = 0x80; mask; mask >>= 1) {
+        while(ARM_DWT_CYCCNT - cyc < CYCLES_400);
+        cyc  = ARM_DWT_CYCCNT;
+        *set = 1;
+        if(pix & mask) {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_400_T1H);
+        } else {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_400_T0H);
+        }
+        *clr = 1;
+      }
+    }
+    while(ARM_DWT_CYCCNT - cyc < CYCLES_400);
+  }
+#endif // NEO_KHZ400
+
+#elif defined(TEENSYDUINO) && (defined(__IMXRT1052__) || defined(__IMXRT1062__))
+#define CYCLES_800_T0H  (F_CPU_ACTUAL / 4000000)
+#define CYCLES_800_T1H  (F_CPU_ACTUAL / 1250000)
+#define CYCLES_800      (F_CPU_ACTUAL /  800000)
+#define CYCLES_400_T0H  (F_CPU_ACTUAL / 2000000)
+#define CYCLES_400_T1H  (F_CPU_ACTUAL /  833333)
+#define CYCLES_400      (F_CPU_ACTUAL /  400000)
+
+  uint8_t           *p   = pixels,
+                    *end = p + numBytes, pix, mask;
+  volatile uint32_t *set = portSetRegister(pin),
+                    *clr = portClearRegister(pin);
+  uint32_t          cyc,
+                    msk = digitalPinToBitMask(pin);
+
+  ARM_DEMCR    |= ARM_DEMCR_TRCENA;
+  ARM_DWT_CTRL |= ARM_DWT_CTRL_CYCCNTENA;
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    cyc = ARM_DWT_CYCCNT + CYCLES_800;
+    while(p < end) {
+      pix = *p++;
+      for(mask = 0x80; mask; mask >>= 1) {
+        while(ARM_DWT_CYCCNT - cyc < CYCLES_800);
+        cyc  = ARM_DWT_CYCCNT;
+        *set = msk;
+        if(pix & mask) {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_800_T1H);
+        } else {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_800_T0H);
+        }
+        *clr = msk;
+      }
+    }
+    while(ARM_DWT_CYCCNT - cyc < CYCLES_800);
+#if defined(NEO_KHZ400)
+  } else { // 400 kHz bitstream
+    cyc = ARM_DWT_CYCCNT + CYCLES_400;
+    while(p < end) {
+      pix = *p++;
+      for(mask = 0x80; mask; mask >>= 1) {
+        while(ARM_DWT_CYCCNT - cyc < CYCLES_400);
+        cyc  = ARM_DWT_CYCCNT;
+        *set = msk;
+        if(pix & mask) {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_400_T1H);
+        } else {
+          while(ARM_DWT_CYCCNT - cyc < CYCLES_400_T0H);
+        }
+        *clr = msk;
+      }
+    }
+    while(ARM_DWT_CYCCNT - cyc < CYCLES_400);
+  }
+#endif // NEO_KHZ400
+
+#elif defined(TEENSYDUINO) && defined(__MKL26Z64__) // Teensy-LC
+
+#if F_CPU == 48000000
+  uint8_t          *p   = pixels,
+                    pix, count, dly,
+                    bitmask = digitalPinToBitMask(pin);
+  volatile uint8_t *reg = portSetRegister(pin);
+  uint32_t          num = numBytes;
+  asm volatile(
+    "L%=_begin:"                      "\n\t"
+     "ldrb  %[pix], [%[p], #0]"       "\n\t"
+     "lsl   %[pix], #24"              "\n\t"
+     "movs  %[count], #7"             "\n\t"
+    "L%=_loop:"                       "\n\t"
+     "lsl   %[pix], #1"               "\n\t"
+     "bcs   L%=_loop_one"             "\n\t"
+     "L%=_loop_zero:"                 "\n\t"
+     "strb  %[bitmask], [%[reg], #0]" "\n\t"
+     "movs  %[dly], #4"               "\n\t"
+    "L%=_loop_delay_T0H:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_loop_delay_T0H"       "\n\t"
+     "strb  %[bitmask], [%[reg], #4]" "\n\t"
+     "movs  %[dly], #13"              "\n\t"
+    "L%=_loop_delay_T0L:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_loop_delay_T0L"       "\n\t"
+     "b     L%=_next"                 "\n\t"
+    "L%=_loop_one:"                   "\n\t"
+     "strb  %[bitmask], [%[reg], #0]" "\n\t"
+     "movs  %[dly], #13"              "\n\t"
+    "L%=_loop_delay_T1H:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_loop_delay_T1H"       "\n\t"
+     "strb  %[bitmask], [%[reg], #4]" "\n\t"
+     "movs  %[dly], #4"               "\n\t"
+    "L%=_loop_delay_T1L:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_loop_delay_T1L"       "\n\t"
+     "nop"                            "\n\t"
+    "L%=_next:"                       "\n\t"
+     "sub   %[count], #1"             "\n\t"
+     "bne   L%=_loop"                 "\n\t"
+     "lsl   %[pix], #1"               "\n\t"
+     "bcs   L%=_last_one"             "\n\t"
+    "L%=_last_zero:"                  "\n\t"
+     "strb  %[bitmask], [%[reg], #0]" "\n\t"
+     "movs  %[dly], #4"               "\n\t"
+    "L%=_last_delay_T0H:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_last_delay_T0H"       "\n\t"
+     "strb  %[bitmask], [%[reg], #4]" "\n\t"
+     "movs  %[dly], #10"              "\n\t"
+    "L%=_last_delay_T0L:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_last_delay_T0L"       "\n\t"
+     "b     L%=_repeat"               "\n\t"
+    "L%=_last_one:"                   "\n\t"
+     "strb  %[bitmask], [%[reg], #0]" "\n\t"
+     "movs  %[dly], #13"              "\n\t"
+    "L%=_last_delay_T1H:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_last_delay_T1H"       "\n\t"
+     "strb  %[bitmask], [%[reg], #4]" "\n\t"
+     "movs  %[dly], #1"               "\n\t"
+    "L%=_last_delay_T1L:"             "\n\t"
+     "sub   %[dly], #1"               "\n\t"
+     "bne   L%=_last_delay_T1L"       "\n\t"
+     "nop"                            "\n\t"
+    "L%=_repeat:"                     "\n\t"
+     "add   %[p], #1"                 "\n\t"
+     "sub   %[num], #1"               "\n\t"
+     "bne   L%=_begin"                "\n\t"
+    "L%=_done:"                       "\n\t"
+    : [p]       "+r"  (p),
+      [pix]     "=&r" (pix),
+      [count]   "=&r" (count),
+      [dly]     "=&r" (dly),
+      [num]     "+r"  (num)
+    : [bitmask] "r"   (bitmask),
+      [reg]     "r"   (reg)
+  );
+#else
+#error "Sorry, only 48 MHz is supported, please set Tools > CPU Speed to 48 MHz"
+#endif // F_CPU == 48000000
+
+// Begin of support for nRF52 based boards  -------------------------
+
+#elif defined(NRF52) || defined(NRF52_SERIES)
+// [[[Begin of the Neopixel NRF52 EasyDMA implementation
+//                                    by the Hackerspace San Salvador]]]
+// This technique uses the PWM peripheral on the NRF52. The PWM uses the
+// EasyDMA feature included on the chip. This technique loads the duty
+// cycle configuration for each cycle when the PWM is enabled. For this
+// to work we need to store a 16 bit configuration for each bit of the
+// RGB(W) values in the pixel buffer.
+// Comparator values for the PWM were hand picked and are guaranteed to
+// be 100% organic to preserve freshness and high accuracy. Current
+// parameters are:
+//   * PWM Clock: 16Mhz
+//   * Minimum step time: 62.5ns
+//   * Time for zero in high (T0H): 0.31ms
+//   * Time for one in high (T1H): 0.75ms
+//   * Cycle time:  1.25us
+//   * Frequency: 800Khz
+// For 400Khz we just double the calculated times.
+// ---------- BEGIN Constants for the EasyDMA implementation -----------
+// The PWM starts the duty cycle in LOW. To start with HIGH we
+// need to set the 15th bit on each register.
+
+// WS2812 (rev A) timing is 0.35 and 0.7us
+//#define MAGIC_T0H               5UL | (0x8000) // 0.3125us
+//#define MAGIC_T1H              12UL | (0x8000) // 0.75us
+
+// WS2812B (rev B) timing is 0.4 and 0.8 us
+#define MAGIC_T0H               6UL | (0x8000) // 0.375us
+#define MAGIC_T1H              13UL | (0x8000) // 0.8125us
+
+// WS2811 (400 khz) timing is 0.5 and 1.2
+#define MAGIC_T0H_400KHz        8UL  | (0x8000) // 0.5us
+#define MAGIC_T1H_400KHz        19UL | (0x8000) // 1.1875us
+
+// For 400Khz, we double value of CTOPVAL
+#define CTOPVAL                20UL            // 1.25us
+#define CTOPVAL_400KHz         40UL            // 2.5us
+
+// ---------- END Constants for the EasyDMA implementation -------------
+//
+// If there is no device available an alternative cycle-counter
+// implementation is tried.
+// The nRF52 runs with a fixed clock of 64Mhz. The alternative
+// implementation is the same as the one used for the Teensy 3.0/1/2 but
+// with the Nordic SDK HAL & registers syntax.
+// The number of cycles was hand picked and is guaranteed to be 100%
+// organic to preserve freshness and high accuracy.
+// ---------- BEGIN Constants for cycle counter implementation ---------
+#define CYCLES_800_T0H  18  // ~0.36 uS
+#define CYCLES_800_T1H  41  // ~0.76 uS
+#define CYCLES_800      71  // ~1.25 uS
+
+#define CYCLES_400_T0H  26  // ~0.50 uS
+#define CYCLES_400_T1H  70  // ~1.26 uS
+#define CYCLES_400      156 // ~2.50 uS
+// ---------- END of Constants for cycle counter implementation --------
+
+  // To support both the SoftDevice + Neopixels we use the EasyDMA
+  // feature from the NRF25. However this technique implies to
+  // generate a pattern and store it on the memory. The actual
+  // memory used in bytes corresponds to the following formula:
+  //              totalMem = numBytes*8*2+(2*2)
+  // The two additional bytes at the end are needed to reset the
+  // sequence.
+  //
+  // If there is not enough memory, we will fall back to cycle counter
+  // using DWT
+  uint32_t  pattern_size   = numBytes*8*sizeof(uint16_t)+2*sizeof(uint16_t);
+  uint16_t* pixels_pattern = NULL;
+
+  NRF_PWM_Type* pwm = NULL;
+
+  // Try to find a free PWM device, which is not enabled
+  // and has no connected pins
+  NRF_PWM_Type* PWM[] = {
+    NRF_PWM0, NRF_PWM1, NRF_PWM2
+#if defined(NRF_PWM3)
+    ,NRF_PWM3
+#endif
+  };
+
+  for(unsigned int device = 0; device < (sizeof(PWM)/sizeof(PWM[0])); device++) {
+    if( (PWM[device]->ENABLE == 0)                            &&
+        (PWM[device]->PSEL.OUT[0] & PWM_PSEL_OUT_CONNECT_Msk) &&
+        (PWM[device]->PSEL.OUT[1] & PWM_PSEL_OUT_CONNECT_Msk) &&
+        (PWM[device]->PSEL.OUT[2] & PWM_PSEL_OUT_CONNECT_Msk) &&
+        (PWM[device]->PSEL.OUT[3] & PWM_PSEL_OUT_CONNECT_Msk)
+    ) {
+      pwm = PWM[device];
+      break;
+    }
+  }
+
+  // only malloc if there is PWM device available
+  if ( pwm != NULL ) {
+    #if defined(ARDUINO_NRF52_ADAFRUIT) // use thread-safe malloc
+      pixels_pattern = (uint16_t *) rtos_malloc(pattern_size);
+    #else
+      pixels_pattern = (uint16_t *) malloc(pattern_size);
+    #endif
+  }
+
+  // Use the identified device to choose the implementation
+  // If a PWM device is available use DMA
+  if( (pixels_pattern != NULL) && (pwm != NULL) ) {
+    uint16_t pos = 0; // bit position
+
+    for(uint16_t n=0; n<numBytes; n++) {
+      uint8_t pix = pixels[n];
+
+      for(uint8_t mask=0x80; mask>0; mask >>= 1) {
+        #if defined(NEO_KHZ400)
+        if( !is800KHz ) {
+          pixels_pattern[pos] = (pix & mask) ? MAGIC_T1H_400KHz : MAGIC_T0H_400KHz;
+        }else
+        #endif
+        {
+          pixels_pattern[pos] = (pix & mask) ? MAGIC_T1H : MAGIC_T0H;
+        }
+
+        pos++;
+      }
+    }
+
+    // Zero padding to indicate the end of que sequence
+    pixels_pattern[pos++] = 0 | (0x8000); // Seq end
+    pixels_pattern[pos++] = 0 | (0x8000); // Seq end
+
+    // Set the wave mode to count UP
+    pwm->MODE = (PWM_MODE_UPDOWN_Up << PWM_MODE_UPDOWN_Pos);
+
+    // Set the PWM to use the 16MHz clock
+    pwm->PRESCALER = (PWM_PRESCALER_PRESCALER_DIV_1 << PWM_PRESCALER_PRESCALER_Pos);
+
+    // Setting of the maximum count
+    // but keeping it on 16Mhz allows for more granularity just
+    // in case someone wants to do more fine-tuning of the timing.
+#if defined(NEO_KHZ400)
+    if( !is800KHz ) {
+      pwm->COUNTERTOP = (CTOPVAL_400KHz << PWM_COUNTERTOP_COUNTERTOP_Pos);
+    }else
+#endif
+    {
+      pwm->COUNTERTOP = (CTOPVAL << PWM_COUNTERTOP_COUNTERTOP_Pos);
+    }
+
+    // Disable loops, we want the sequence to repeat only once
+    pwm->LOOP = (PWM_LOOP_CNT_Disabled << PWM_LOOP_CNT_Pos);
+
+    // On the "Common" setting the PWM uses the same pattern for the
+    // for supported sequences. The pattern is stored on half-word
+    // of 16bits
+    pwm->DECODER = (PWM_DECODER_LOAD_Common << PWM_DECODER_LOAD_Pos) |
+                   (PWM_DECODER_MODE_RefreshCount << PWM_DECODER_MODE_Pos);
+
+    // Pointer to the memory storing the patter
+    pwm->SEQ[0].PTR = (uint32_t)(pixels_pattern) << PWM_SEQ_PTR_PTR_Pos;
+
+    // Calculation of the number of steps loaded from memory.
+    pwm->SEQ[0].CNT = (pattern_size/sizeof(uint16_t)) << PWM_SEQ_CNT_CNT_Pos;
+
+    // The following settings are ignored with the current config.
+    pwm->SEQ[0].REFRESH  = 0;
+    pwm->SEQ[0].ENDDELAY = 0;
+
+    // The Neopixel implementation is a blocking algorithm. DMA
+    // allows for non-blocking operation. To "simulate" a blocking
+    // operation we enable the interruption for the end of sequence
+    // and block the execution thread until the event flag is set by
+    // the peripheral.
+//    pwm->INTEN |= (PWM_INTEN_SEQEND0_Enabled<<PWM_INTEN_SEQEND0_Pos);
+
+    // PSEL must be configured before enabling PWM
+    #if defined(ARDUINO_ARCH_NRF52840)
+    pwm->PSEL.OUT[0] = g_APinDescription[pin].name;
+    #else
+    pwm->PSEL.OUT[0] = g_ADigitalPinMap[pin];
+    #endif
+
+    // Enable the PWM
+    pwm->ENABLE = 1;
+
+    // After all of this and many hours of reading the documentation
+    // we are ready to start the sequence...
+    pwm->EVENTS_SEQEND[0]  = 0;
+    pwm->TASKS_SEQSTART[0] = 1;
+
+    // But we have to wait for the flag to be set.
+    while(!pwm->EVENTS_SEQEND[0])
+    {
+      #if defined(ARDUINO_NRF52_ADAFRUIT) || defined(ARDUINO_ARCH_NRF52840)
+      yield();
+      #endif
+    }
+
+    // Before leave we clear the flag for the event.
+    pwm->EVENTS_SEQEND[0] = 0;
+
+    // We need to disable the device and disconnect
+    // all the outputs before leave or the device will not
+    // be selected on the next call.
+    // TODO: Check if disabling the device causes performance issues.
+    pwm->ENABLE = 0;
+
+    pwm->PSEL.OUT[0] = 0xFFFFFFFFUL;
+
+    #if defined(ARDUINO_NRF52_ADAFRUIT)  // use thread-safe free
+      rtos_free(pixels_pattern);
+    #else
+      free(pixels_pattern);
+    #endif
+  }// End of DMA implementation
+  // ---------------------------------------------------------------------
+  else{
+#ifndef ARDUINO_ARCH_NRF52840     
+    // Fall back to DWT
+    #if defined(ARDUINO_NRF52_ADAFRUIT)
+      // Bluefruit Feather 52 uses freeRTOS
+      // Critical Section is used since it does not block SoftDevice execution
+      taskENTER_CRITICAL();
+    #elif defined(NRF52_DISABLE_INT)
+      // If you are using the Bluetooth SoftDevice we advise you to not disable
+      // the interrupts. Disabling the interrupts even for short periods of time
+      // causes the SoftDevice to stop working.
+      // Disable the interrupts only in cases where you need high performance for
+      // the LEDs and if you are not using the EasyDMA feature.
+      __disable_irq();
+    #endif
+
+    NRF_GPIO_Type* nrf_port = (NRF_GPIO_Type*) digitalPinToPort(pin);
+    uint32_t pinMask = digitalPinToBitMask(pin);
+
+    uint32_t CYCLES_X00     = CYCLES_800;
+    uint32_t CYCLES_X00_T1H = CYCLES_800_T1H;
+    uint32_t CYCLES_X00_T0H = CYCLES_800_T0H;
+
+#if defined(NEO_KHZ400)
+    if( !is800KHz )
+    {
+      CYCLES_X00     = CYCLES_400;
+      CYCLES_X00_T1H = CYCLES_400_T1H;
+      CYCLES_X00_T0H = CYCLES_400_T0H;
+    }
+#endif
+
+    // Enable DWT in debug core
+    CoreDebug->DEMCR |= CoreDebug_DEMCR_TRCENA_Msk;
+    DWT->CTRL |= DWT_CTRL_CYCCNTENA_Msk;
+
+    // Tries to re-send the frame if is interrupted by the SoftDevice.
+    while(1) {
+      uint8_t *p = pixels;
+
+      uint32_t cycStart = DWT->CYCCNT;
+      uint32_t cyc = 0;
+
+      for(uint16_t n=0; n<numBytes; n++) {
+        uint8_t pix = *p++;
+
+        for(uint8_t mask = 0x80; mask; mask >>= 1) {
+          while(DWT->CYCCNT - cyc < CYCLES_X00);
+          cyc  = DWT->CYCCNT;
+
+          nrf_port->OUTSET |= pinMask;
+
+          if(pix & mask) {
+            while(DWT->CYCCNT - cyc < CYCLES_X00_T1H);
+          } else {
+            while(DWT->CYCCNT - cyc < CYCLES_X00_T0H);
+          }
+
+          nrf_port->OUTCLR |= pinMask;
+        }
+      }
+      while(DWT->CYCCNT - cyc < CYCLES_X00);
+
+
+      // If total time longer than 25%, resend the whole data.
+      // Since we are likely to be interrupted by SoftDevice
+      if ( (DWT->CYCCNT - cycStart) < ( 8*numBytes*((CYCLES_X00*5)/4) ) ) {
+        break;
+      }
+
+      // re-send need 300us delay
+      delayMicroseconds(300);
+    }
+
+    // Enable interrupts again
+    #if defined(ARDUINO_NRF52_ADAFRUIT)
+      taskEXIT_CRITICAL();
+    #elif defined(NRF52_DISABLE_INT)
+      __enable_irq();
+    #endif
+#endif
+  }
+// END of NRF52 implementation
+
+#elif defined (__SAMD21E17A__) || defined(__SAMD21G18A__)  || defined(__SAMD21E18A__) || defined(__SAMD21J18A__) // Arduino Zero, Gemma/Trinket M0, SODAQ Autonomo and others
+  // Tried this with a timer/counter, couldn't quite get adequate
+  // resolution. So yay, you get a load of goofball NOPs...
+
+  uint8_t  *ptr, *end, p, bitMask, portNum;
+  uint32_t  pinMask;
+
+  portNum =  g_APinDescription[pin].ulPort;
+  pinMask =  1ul << g_APinDescription[pin].ulPin;
+  ptr     =  pixels;
+  end     =  ptr + numBytes;
+  p       = *ptr++;
+  bitMask =  0x80;
+
+  volatile uint32_t *set = &(PORT->Group[portNum].OUTSET.reg),
+                    *clr = &(PORT->Group[portNum].OUTCLR.reg);
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    for(;;) {
+      *set = pinMask;
+      asm("nop; nop; nop; nop; nop; nop; nop; nop;");
+      if(p & bitMask) {
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop;");
+        *clr = pinMask;
+      } else {
+        *clr = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop;");
+      }
+      if(bitMask >>= 1) {
+        asm("nop; nop; nop; nop; nop; nop; nop; nop; nop;");
+      } else {
+        if(ptr >= end) break;
+        p       = *ptr++;
+        bitMask = 0x80;
+      }
+    }
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz bitstream
+    for(;;) {
+      *set = pinMask;
+      asm("nop; nop; nop; nop; nop; nop; nop; nop; nop; nop; nop;");
+      if(p & bitMask) {
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop;");
+        *clr = pinMask;
+      } else {
+        *clr = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop;");
+      }
+      asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+          "nop; nop; nop; nop; nop; nop; nop; nop;"
+          "nop; nop; nop; nop; nop; nop; nop; nop;"
+          "nop; nop; nop; nop; nop; nop; nop; nop;");
+      if(bitMask >>= 1) {
+        asm("nop; nop; nop; nop; nop; nop; nop;");
+      } else {
+        if(ptr >= end) break;
+        p       = *ptr++;
+        bitMask = 0x80;
+      }
+    }
+  }
+#endif
+
+#elif defined (__SAMD51__) // M4
+
+  uint8_t  *ptr, *end, p, bitMask, portNum, bit;
+  uint32_t  pinMask;
+
+  portNum =  g_APinDescription[pin].ulPort;
+  pinMask =  1ul << g_APinDescription[pin].ulPin;
+  ptr     =  pixels;
+  end     =  ptr + numBytes;
+  p       = *ptr++;
+  bitMask =  0x80;
+
+  volatile uint32_t *set = &(PORT->Group[portNum].OUTSET.reg),
+                    *clr = &(PORT->Group[portNum].OUTCLR.reg);
+
+  // SAMD51 overclock-compatible timing is only a mild abomination.
+  // It uses SysTick for a consistent clock reference regardless of
+  // optimization / cache settings.  That's the good news.  The bad news,
+  // since SysTick->VAL is a volatile type it's slow to access...and then,
+  // with the SysTick interval that Arduino sets up (1 ms), this would
+  // require a subtract and MOD operation for gauging elapsed time, and
+  // all taken in combination that lacks adequate temporal resolution
+  // for NeoPixel timing.  So a kind of horrible thing is done here...
+  // since interrupts are turned off anyway and it's generally accepted
+  // by now that we're gonna lose track of time in the NeoPixel lib,
+  // the SysTick timer is reconfigured for a period matching the NeoPixel
+  // bit timing (either 800 or 400 KHz) and we watch SysTick->VAL very
+  // closely (just a threshold, no subtract or MOD or anything) and that
+  // seems to work just well enough.  When finished, the SysTick
+  // peripheral is set back to its original state.
+
+  uint32_t t0, t1, top, ticks,
+           saveLoad = SysTick->LOAD, saveVal = SysTick->VAL;
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    top =       (uint32_t)(F_CPU * 0.00000125); // Bit hi + lo = 1.25 uS
+    t0  = top - (uint32_t)(F_CPU * 0.00000040); // 0 = 0.4 uS hi
+    t1  = top - (uint32_t)(F_CPU * 0.00000080); // 1 = 0.8 uS hi
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz bitstream
+    top =       (uint32_t)(F_CPU * 0.00000250); // Bit hi + lo = 2.5 uS
+    t0  = top - (uint32_t)(F_CPU * 0.00000050); // 0 = 0.5 uS hi
+    t1  = top - (uint32_t)(F_CPU * 0.00000120); // 1 = 1.2 uS hi
+  }
+#endif
+
+  SysTick->LOAD = top;               // Config SysTick for NeoPixel bit freq
+  SysTick->VAL  = top;               // Set to start value (counts down)
+  (void)SysTick->VAL;                // Dummy read helps sync up 1st bit
+
+  for(;;) {
+    *set  = pinMask;                 // Set output high
+    ticks = (p & bitMask) ? t1 : t0; // SysTick threshold,
+    while(SysTick->VAL > ticks);     // wait for it
+    *clr  = pinMask;                 // Set output low
+    if(!(bitMask >>= 1)) {           // Next bit for this byte...done?
+      if(ptr >= end) break;          // If last byte sent, exit loop
+      p       = *ptr++;              // Fetch next byte
+      bitMask = 0x80;                // Reset bitmask
+    }
+    while(SysTick->VAL <= ticks);    // Wait for rollover to 'top'
+  }
+
+  SysTick->LOAD = saveLoad;          // Restore SysTick rollover to 1 ms
+  SysTick->VAL  = saveVal;           // Restore SysTick value
+
+#elif defined (ARDUINO_STM32_FEATHER) // FEATHER WICED (120MHz)
+
+  // Tried this with a timer/counter, couldn't quite get adequate
+  // resolution. So yay, you get a load of goofball NOPs...
+
+  uint8_t  *ptr, *end, p, bitMask;
+  uint32_t  pinMask;
+
+  pinMask =  BIT(PIN_MAP[pin].gpio_bit);
+  ptr     =  pixels;
+  end     =  ptr + numBytes;
+  p       = *ptr++;
+  bitMask =  0x80;
+
+  volatile uint16_t *set = &(PIN_MAP[pin].gpio_device->regs->BSRRL);
+  volatile uint16_t *clr = &(PIN_MAP[pin].gpio_device->regs->BSRRH);
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    for(;;) {
+      if(p & bitMask) { // ONE
+        // High 800ns
+        *set = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop;");
+        // Low 450ns
+        *clr = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop;");
+      } else { // ZERO
+        // High 400ns
+        *set = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop;");
+        // Low 850ns
+        *clr = pinMask;
+        asm("nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop; nop; nop; nop; nop;"
+            "nop; nop; nop; nop;");
+      }
+      if(bitMask >>= 1) {
+        // Move on to the next pixel
+        asm("nop;");
+      } else {
+        if(ptr >= end) break;
+        p       = *ptr++;
+        bitMask = 0x80;
+      }
+    }
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz bitstream
+    // ToDo!
+  }
+#endif
+
+#elif defined(TARGET_LPC1768)
+  uint8_t  *ptr, *end, p, bitMask;
+  ptr     =  pixels;
+  end     =  ptr + numBytes;
+  p       = *ptr++;
+  bitMask =  0x80;
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    for(;;) {
+      if(p & bitMask) {
+        // data ONE high
+        // min: 550 typ: 700 max: 5,500
+        gpio_set(pin);
+        time::delay_ns(550);
+        // min: 450 typ: 600 max: 5,000
+        gpio_clear(pin);
+        time::delay_ns(450);
+      } else {
+        // data ZERO high
+        // min: 200  typ: 350 max: 500
+        gpio_set(pin);
+        time::delay_ns(200);
+        // data low
+        // min: 450 typ: 600 max: 5,000
+        gpio_clear(pin);
+        time::delay_ns(450);
+      }
+      if(bitMask >>= 1) {
+        // Move on to the next pixel
+        asm("nop;");
+      } else {
+        if(ptr >= end) break;
+        p       = *ptr++;
+        bitMask = 0x80;
+      }
+    }
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz bitstream
+    // ToDo!
+  }
+#endif
+#elif defined(ARDUINO_ARCH_STM32) || defined(ARDUINO_ARCH_ARDUINO_CORE_STM32)
+  uint8_t           *p   = pixels, *end = p + numBytes,
+                    pix = *p++, mask = 0x80;
+  uint32_t          cyc;
+  uint32_t saveLoad = SysTick->LOAD, saveVal = SysTick->VAL;
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    uint32_t top = (F_CPU /  800000);       // 1.25µs
+    uint32_t t0  = top - (F_CPU / 2500000); // 0.4µs
+    uint32_t t1  = top - (F_CPU / 1250000); // 0.8µs
+    SysTick->LOAD = top - 1; // Config SysTick for NeoPixel bit freq
+    SysTick->VAL  = 0; // Set to start value
+    for(;;) {
+      LL_GPIO_SetOutputPin(gpioPort, gpioPin);
+      cyc = (pix & mask) ? t1 : t0;
+      while(SysTick->VAL > cyc);
+      LL_GPIO_ResetOutputPin(gpioPort, gpioPin);
+      if(!(mask >>= 1)) {
+        if(p >= end) break;
+        pix       = *p++;
+        mask = 0x80;
+      }
+      while(SysTick->VAL <= cyc);
+    }
+#if defined(NEO_KHZ400)
+  } else { // 400 kHz bitstream
+    uint32_t top = (F_CPU /  400000);       // 2.5µs
+    uint32_t t0  = top - (F_CPU / 2000000); // 0.5µs
+    uint32_t t1  = top - (F_CPU /  833333); // 1.2µs
+    SysTick->LOAD = top - 1; // Config SysTick for NeoPixel bit freq
+    SysTick->VAL  = 0;       // Set to start value
+    for(;;) {
+      LL_GPIO_SetOutputPin(gpioPort, gpioPin);
+      cyc = (pix & mask) ? t1 : t0;
+      while(SysTick->VAL > cyc);
+      LL_GPIO_ResetOutputPin(gpioPort, gpioPin);
+      if(!(mask >>= 1)) {
+        if(p >= end) break;
+        pix       = *p++;
+        mask = 0x80;
+      }
+      while(SysTick->VAL <= cyc);
+    }
+  }
+#endif // NEO_KHZ400
+  SysTick->LOAD = saveLoad;          // Restore SysTick rollover to 1 ms
+  SysTick->VAL  = saveVal;           // Restore SysTick value
+#elif defined (NRF51)
+  uint8_t          *p   = pixels,
+                    pix, count, mask;
+  int32_t         num = numBytes;
+  unsigned int bitmask = ( 1 << g_ADigitalPinMap[pin] );
+// https://github.com/sandeepmistry/arduino-nRF5/blob/dc53980c8bac27898fca90d8ecb268e11111edc1/variants/BBCmicrobit/variant.cpp
+
+  volatile unsigned int *reg = (unsigned int *) (0x50000000UL + 0x508);
+
+// https://github.com/sandeepmistry/arduino-nRF5/blob/dc53980c8bac27898fca90d8ecb268e11111edc1/cores/nRF5/SDK/components/device/nrf51.h
+// http://www.iot-programmer.com/index.php/books/27-micro-bit-iot-in-c/chapters-micro-bit-iot-in-c/47-micro-bit-iot-in-c-fast-memory-mapped-gpio?showall=1
+// https://github.com/Microsoft/pxt-neopixel/blob/master/sendbuffer.asm
+
+  asm volatile(
+    // "cpsid i" ; disable irq
+
+    //    b .start
+    "b  L%=_start"                    "\n\t"
+    // .nextbit:               ;            C0
+    "L%=_nextbit:"                    "\n\t"          //;            C0
+    //    str r1, [r3, #0]    ; pin := hi  C2
+    "strb %[bitmask], [%[reg], #0]"   "\n\t"          //; pin := hi  C2
+    //    tst r6, r0          ;            C3
+    "tst %[mask], %[pix]"             "\n\t"//          ;            C3
+    //    bne .islate         ;            C4
+    "bne L%=_islate"                  "\n\t"          //;            C4
+    //    str r1, [r2, #0]    ; pin := lo  C6
+    "strb %[bitmask], [%[reg], #4]"   "\n\t"          //; pin := lo  C6
+    // .islate:
+    "L%=_islate:"                     "\n\t"
+    //    lsrs r6, r6, #1     ; r6 >>= 1   C7
+    "lsr %[mask], %[mask], #1"       "\n\t"          //; r6 >>= 1   C7
+    //    bne .justbit        ;            C8
+    "bne L%=_justbit"                 "\n\t"          //;            C8
+
+    //    ; not just a bit - need new byte
+    //    adds r4, #1         ; r4++       C9
+    "add %[p], #1"                   "\n\t"          //; r4++       C9
+    //    subs r5, #1         ; r5--       C10
+    "sub %[num], #1"                 "\n\t"          //; r5--       C10
+    //    bcc .stop           ; if (r5<0) goto .stop  C11
+    "bcc L%=_stop"                    "\n\t"          //; if (r5<0) goto .stop  C11
+    // .start:
+    "L%=_start:"
+    //    movs r6, #0x80      ; reset mask C12
+    "movs %[mask], #0x80"             "\n\t"          //; reset mask C12
+    //    nop                 ;            C13
+    "nop"                             "\n\t"          //;            C13
+
+    // .common:               ;             C13
+    "L%=_common:"                     "\n\t"          //;            C13
+    //    str r1, [r2, #0]   ; pin := lo   C15
+    "strb %[bitmask], [%[reg], #4]"   "\n\t"          //; pin := lo  C15
+    //    ; always re-load byte - it just fits with the cycles better this way
+    //    ldrb r0, [r4, #0]  ; r0 := *r4   C17
+    "ldrb  %[pix], [%[p], #0]"        "\n\t"          //; r0 := *r4   C17
+    //    b .nextbit         ;             C20
+    "b L%=_nextbit"                   "\n\t"          //;             C20
+
+    // .justbit: ; C10
+    "L%=_justbit:"                    "\n\t"          //; C10
+    //    ; no nops, branch taken is already 3 cycles
+    //    b .common ; C13
+    "b L%=_common"                    "\n\t"          //; C13
+
+    // .stop:
+    "L%=_stop:"                       "\n\t"
+    //    str r1, [r2, #0]   ; pin := lo
+    "strb %[bitmask], [%[reg], #4]"   "\n\t"          //; pin := lo
+    //    cpsie i            ; enable irq
+
+    : [p] "+r" (p),
+    [pix] "=&r" (pix),
+    [count] "=&r" (count),
+    [mask] "=&r" (mask),
+    [num] "+r" (num)
+    : [bitmask] "r" (bitmask),
+    [reg] "r" (reg)
+  );
+
+#elif defined(__SAM3X8E__) // Arduino Due
+
+  #define SCALE      VARIANT_MCK / 2UL / 1000000UL
+  #define INST       (2UL * F_CPU / VARIANT_MCK)
+  #define TIME_800_0 ((int)(0.40 * SCALE + 0.5) - (5 * INST))
+  #define TIME_800_1 ((int)(0.80 * SCALE + 0.5) - (5 * INST))
+  #define PERIOD_800 ((int)(1.25 * SCALE + 0.5) - (5 * INST))
+  #define TIME_400_0 ((int)(0.50 * SCALE + 0.5) - (5 * INST))
+  #define TIME_400_1 ((int)(1.20 * SCALE + 0.5) - (5 * INST))
+  #define PERIOD_400 ((int)(2.50 * SCALE + 0.5) - (5 * INST))
+
+  int             pinMask, time0, time1, period, t;
+  Pio            *port;
+  volatile WoReg *portSet, *portClear, *timeValue, *timeReset;
+  uint8_t        *p, *end, pix, mask;
+
+  pmc_set_writeprotect(false);
+  pmc_enable_periph_clk((uint32_t)TC3_IRQn);
+  TC_Configure(TC1, 0,
+    TC_CMR_WAVE | TC_CMR_WAVSEL_UP | TC_CMR_TCCLKS_TIMER_CLOCK1);
+  TC_Start(TC1, 0);
+
+  pinMask   = g_APinDescription[pin].ulPin; // Don't 'optimize' these into
+  port      = g_APinDescription[pin].pPort; // declarations above. Want to
+  portSet   = &(port->PIO_SODR);            // burn a few cycles after
+  portClear = &(port->PIO_CODR);            // starting timer to minimize
+  timeValue = &(TC1->TC_CHANNEL[0].TC_CV);  // the initial 'while'.
+  timeReset = &(TC1->TC_CHANNEL[0].TC_CCR);
+  p         =  pixels;
+  end       =  p + numBytes;
+  pix       = *p++;
+  mask      = 0x80;
+
+#if defined(NEO_KHZ400) // 800 KHz check needed only if 400 KHz support enabled
+  if(is800KHz) {
+#endif
+    time0  = TIME_800_0;
+    time1  = TIME_800_1;
+    period = PERIOD_800;
+#if defined(NEO_KHZ400)
+  } else { // 400 KHz bitstream
+    time0  = TIME_400_0;
+    time1  = TIME_400_1;
+    period = PERIOD_400;
+  }
+#endif
+
+  for(t = time0;; t = time0) {
+    if(pix & mask) t = time1;
+    while(*timeValue < (unsigned)period);
+    *portSet   = pinMask;
+    *timeReset = TC_CCR_CLKEN | TC_CCR_SWTRG;
+    while(*timeValue < (unsigned)t);
+    *portClear = pinMask;
+    if(!(mask >>= 1)) {   // This 'inside-out' loop logic utilizes
+      if(p >= end) break; // idle time to minimize inter-byte delays.
+      pix = *p++;
+      mask = 0x80;
+    }
+  }
+  while(*timeValue < (unsigned)period); // Wait for last bit
+  TC_Stop(TC1, 0);
+
+#endif // end Due
+
+// END ARM ----------------------------------------------------------------
+
+
+#elif defined(ESP8266) || defined(ESP32)
+
+// ESP8266 ----------------------------------------------------------------
+
+  // ESP8266 show() is external to enforce ICACHE_RAM_ATTR execution
+  espShow(pin, pixels, numBytes, is800KHz);
+
+#elif defined(KENDRYTE_K210)
+
+  k210Show(pin, pixels, numBytes, is800KHz);
+  
+#elif defined(__ARDUINO_ARC__)
+
+// Arduino 101  -----------------------------------------------------------
+
+#define NOPx7 { __builtin_arc_nop(); \
+  __builtin_arc_nop(); __builtin_arc_nop(); \
+  __builtin_arc_nop(); __builtin_arc_nop(); \
+  __builtin_arc_nop(); __builtin_arc_nop(); }
+
+  PinDescription *pindesc = &g_APinDescription[pin];
+  register uint32_t loop = 8 * numBytes; // one loop to handle all bytes and all bits
+  register uint8_t *p = pixels;
+  register uint32_t currByte = (uint32_t) (*p);
+  register uint32_t currBit = 0x80 & currByte;
+  register uint32_t bitCounter = 0;
+  register uint32_t first = 1;
+
+  // The loop is unusual. Very first iteration puts all the way LOW to the wire -
+  // constant LOW does not affect NEOPIXEL, so there is no visible effect displayed.
+  // During that very first iteration CPU caches instructions in the loop.
+  // Because of the caching process, "CPU slows down". NEOPIXEL pulse is very time sensitive
+  // that's why we let the CPU cache first and we start regular pulse from 2nd iteration
+  if (pindesc->ulGPIOType == SS_GPIO) {
+    register uint32_t reg = pindesc->ulGPIOBase + SS_GPIO_SWPORTA_DR;
+    uint32_t reg_val = __builtin_arc_lr((volatile uint32_t)reg);
+    register uint32_t reg_bit_high = reg_val | (1 << pindesc->ulGPIOId);
+    register uint32_t reg_bit_low  = reg_val & ~(1 << pindesc->ulGPIOId);
+
+    loop += 1; // include first, special iteration
+    while(loop--) {
+      if(!first) {
+        currByte <<= 1;
+        bitCounter++;
+      }
+
+      // 1 is >550ns high and >450ns low; 0 is 200..500ns high and >450ns low
+      __builtin_arc_sr(first ? reg_bit_low : reg_bit_high, (volatile uint32_t)reg);
+      if(currBit) { // ~400ns HIGH (740ns overall)
+        NOPx7
+        NOPx7
+      }
+      // ~340ns HIGH
+      NOPx7
+     __builtin_arc_nop();
+
+      // 820ns LOW; per spec, max allowed low here is 5000ns */
+      __builtin_arc_sr(reg_bit_low, (volatile uint32_t)reg);
+      NOPx7
+      NOPx7
+
+      if(bitCounter >= 8) {
+        bitCounter = 0;
+        currByte = (uint32_t) (*++p);
+      }
+
+      currBit = 0x80 & currByte;
+      first = 0;
+    }
+  } else if(pindesc->ulGPIOType == SOC_GPIO) {
+    register uint32_t reg = pindesc->ulGPIOBase + SOC_GPIO_SWPORTA_DR;
+    uint32_t reg_val = MMIO_REG_VAL(reg);
+    register uint32_t reg_bit_high = reg_val | (1 << pindesc->ulGPIOId);
+    register uint32_t reg_bit_low  = reg_val & ~(1 << pindesc->ulGPIOId);
+
+    loop += 1; // include first, special iteration
+    while(loop--) {
+      if(!first) {
+        currByte <<= 1;
+        bitCounter++;
+      }
+      MMIO_REG_VAL(reg) = first ? reg_bit_low : reg_bit_high;
+      if(currBit) { // ~430ns HIGH (740ns overall)
+        NOPx7
+        NOPx7
+        __builtin_arc_nop();
+      }
+      // ~310ns HIGH
+      NOPx7
+
+      // 850ns LOW; per spec, max allowed low here is 5000ns */
+      MMIO_REG_VAL(reg) = reg_bit_low;
+      NOPx7
+      NOPx7
+
+      if(bitCounter >= 8) {
+        bitCounter = 0;
+        currByte = (uint32_t) (*++p);
+      }
+
+      currBit = 0x80 & currByte;
+      first = 0;
+    }
+  }
+
+#else
+#error Architecture not supported
+#endif
+
+
+// END ARCHITECTURE SELECT ------------------------------------------------
+
+#if !( defined(NRF52) || defined(NRF52_SERIES) )
+  interrupts();
+#endif
+
+  endTime = micros(); // Save EOD time for latch on next call
+}
+
+/*!
+  @brief   Set/change the NeoPixel output pin number. Previous pin,
+           if any, is set to INPUT and the new pin is set to OUTPUT.
+  @param   p  Arduino pin number (-1 = no pin).
+*/
+void Adafruit_NeoPixel::setPin(int16_t p) {
+  if(begun && (pin >= 0)) pinMode(pin, INPUT); // Disable existing out pin
+  pin = p;
+  if(begun) {
+    pinMode(p, OUTPUT);
+    digitalWrite(p, LOW);
+  }
+#if defined(__AVR__)
+  port    = portOutputRegister(digitalPinToPort(p));
+  pinMask = digitalPinToBitMask(p);
+#endif
+#if defined(ARDUINO_ARCH_STM32) || defined(ARDUINO_ARCH_ARDUINO_CORE_STM32)
+  gpioPort = digitalPinToPort(p);
+  gpioPin = STM_LL_GPIO_PIN(digitalPinToPinName(p));
+#endif
+}
+
+/*!
+  @brief   Set a pixel's color using separate red, green and blue
+           components. If using RGBW pixels, white will be set to 0.
+  @param   n  Pixel index, starting from 0.
+  @param   r  Red brightness, 0 = minimum (off), 255 = maximum.
+  @param   g  Green brightness, 0 = minimum (off), 255 = maximum.
+  @param   b  Blue brightness, 0 = minimum (off), 255 = maximum.
+*/
+void Adafruit_NeoPixel::setPixelColor(
+ uint16_t n, uint8_t r, uint8_t g, uint8_t b) {
+
+  if(n < numLEDs) {
+    if(brightness) { // See notes in setBrightness()
+      r = (r * brightness) >> 8;
+      g = (g * brightness) >> 8;
+      b = (b * brightness) >> 8;
+    }
+    uint8_t *p;
+    if(wOffset == rOffset) { // Is an RGB-type strip
+      p = &pixels[n * 3];    // 3 bytes per pixel
+    } else {                 // Is a WRGB-type strip
+      p = &pixels[n * 4];    // 4 bytes per pixel
+      p[wOffset] = 0;        // But only R,G,B passed -- set W to 0
+    }
+    p[rOffset] = r;          // R,G,B always stored
+    p[gOffset] = g;
+    p[bOffset] = b;
+  }
+}
+
+/*!
+  @brief   Set a pixel's color using separate red, green, blue and white
+           components (for RGBW NeoPixels only).
+  @param   n  Pixel index, starting from 0.
+  @param   r  Red brightness, 0 = minimum (off), 255 = maximum.
+  @param   g  Green brightness, 0 = minimum (off), 255 = maximum.
+  @param   b  Blue brightness, 0 = minimum (off), 255 = maximum.
+  @param   w  White brightness, 0 = minimum (off), 255 = maximum, ignored
+              if using RGB pixels.
+*/
+void Adafruit_NeoPixel::setPixelColor(
+ uint16_t n, uint8_t r, uint8_t g, uint8_t b, uint8_t w) {
+
+  if(n < numLEDs) {
+    if(brightness) { // See notes in setBrightness()
+      r = (r * brightness) >> 8;
+      g = (g * brightness) >> 8;
+      b = (b * brightness) >> 8;
+      w = (w * brightness) >> 8;
+    }
+    uint8_t *p;
+    if(wOffset == rOffset) { // Is an RGB-type strip
+      p = &pixels[n * 3];    // 3 bytes per pixel (ignore W)
+    } else {                 // Is a WRGB-type strip
+      p = &pixels[n * 4];    // 4 bytes per pixel
+      p[wOffset] = w;        // Store W
+    }
+    p[rOffset] = r;          // Store R,G,B
+    p[gOffset] = g;
+    p[bOffset] = b;
+  }
+}
+
+/*!
+  @brief   Set a pixel's color using a 32-bit 'packed' RGB or RGBW value.
+  @param   n  Pixel index, starting from 0.
+  @param   c  32-bit color value. Most significant byte is white (for RGBW
+              pixels) or ignored (for RGB pixels), next is red, then green,
+              and least significant byte is blue.
+*/
+void Adafruit_NeoPixel::setPixelColor(uint16_t n, uint32_t c) {
+  if(n < numLEDs) {
+    uint8_t *p,
+      r = (uint8_t)(c >> 16),
+      g = (uint8_t)(c >>  8),
+      b = (uint8_t)c;
+    if(brightness) { // See notes in setBrightness()
+      r = (r * brightness) >> 8;
+      g = (g * brightness) >> 8;
+      b = (b * brightness) >> 8;
+    }
+    if(wOffset == rOffset) {
+      p = &pixels[n * 3];
+    } else {
+      p = &pixels[n * 4];
+      uint8_t w = (uint8_t)(c >> 24);
+      p[wOffset] = brightness ? ((w * brightness) >> 8) : w;
+    }
+    p[rOffset] = r;
+    p[gOffset] = g;
+    p[bOffset] = b;
+  }
+}
+
+/*!
+  @brief   Fill all or part of the NeoPixel strip with a color.
+  @param   c      32-bit color value. Most significant byte is white (for
+                  RGBW pixels) or ignored (for RGB pixels), next is red,
+                  then green, and least significant byte is blue. If all
+                  arguments are unspecified, this will be 0 (off).
+  @param   first  Index of first pixel to fill, starting from 0. Must be
+                  in-bounds, no clipping is performed. 0 if unspecified.
+  @param   count  Number of pixels to fill, as a positive value. Passing
+                  0 or leaving unspecified will fill to end of strip.
+*/
+void Adafruit_NeoPixel::fill(uint32_t c, uint16_t first, uint16_t count) {
+  uint16_t i, end;
+
+  if(first >= numLEDs) {
+    return; // If first LED is past end of strip, nothing to do
+  }
+
+  // Calculate the index ONE AFTER the last pixel to fill
+  if(count == 0) {
+    // Fill to end of strip
+    end = numLEDs;
+  } else {
+    // Ensure that the loop won't go past the last pixel
+    end = first + count;
+    if(end > numLEDs) end = numLEDs;
+  }
+
+  for(i = first; i < end; i++) {
+    this->setPixelColor(i, c);
+  }
+}
+
+/*!
+  @brief   Convert hue, saturation and value into a packed 32-bit RGB color
+           that can be passed to setPixelColor() or other RGB-compatible
+           functions.
+  @param   hue  An unsigned 16-bit value, 0 to 65535, representing one full
+                loop of the color wheel, which allows 16-bit hues to "roll
+                over" while still doing the expected thing (and allowing
+                more precision than the wheel() function that was common to
+                prior NeoPixel examples).
+  @param   sat  Saturation, 8-bit value, 0 (min or pure grayscale) to 255
+                (max or pure hue). Default of 255 if unspecified.
+  @param   val  Value (brightness), 8-bit value, 0 (min / black / off) to
+                255 (max or full brightness). Default of 255 if unspecified.
+  @return  Packed 32-bit RGB with the most significant byte set to 0 -- the
+           white element of WRGB pixels is NOT utilized. Result is linearly
+           but not perceptually correct, so you may want to pass the result
+           through the gamma32() function (or your own gamma-correction
+           operation) else colors may appear washed out. This is not done
+           automatically by this function because coders may desire a more
+           refined gamma-correction function than the simplified
+           one-size-fits-all operation of gamma32(). Diffusing the LEDs also
+           really seems to help when using low-saturation colors.
+*/
+uint32_t Adafruit_NeoPixel::ColorHSV(uint16_t hue, uint8_t sat, uint8_t val) {
+
+  uint8_t r, g, b;
+
+  // Remap 0-65535 to 0-1529. Pure red is CENTERED on the 64K rollover;
+  // 0 is not the start of pure red, but the midpoint...a few values above
+  // zero and a few below 65536 all yield pure red (similarly, 32768 is the
+  // midpoint, not start, of pure cyan). The 8-bit RGB hexcone (256 values
+  // each for red, green, blue) really only allows for 1530 distinct hues
+  // (not 1536, more on that below), but the full unsigned 16-bit type was
+  // chosen for hue so that one's code can easily handle a contiguous color
+  // wheel by allowing hue to roll over in either direction.
+  hue = (hue * 1530L + 32768) / 65536;
+  // Because red is centered on the rollover point (the +32768 above,
+  // essentially a fixed-point +0.5), the above actually yields 0 to 1530,
+  // where 0 and 1530 would yield the same thing. Rather than apply a
+  // costly modulo operator, 1530 is handled as a special case below.
+
+  // So you'd think that the color "hexcone" (the thing that ramps from
+  // pure red, to pure yellow, to pure green and so forth back to red,
+  // yielding six slices), and with each color component having 256
+  // possible values (0-255), might have 1536 possible items (6*256),
+  // but in reality there's 1530. This is because the last element in
+  // each 256-element slice is equal to the first element of the next
+  // slice, and keeping those in there this would create small
+  // discontinuities in the color wheel. So the last element of each
+  // slice is dropped...we regard only elements 0-254, with item 255
+  // being picked up as element 0 of the next slice. Like this:
+  // Red to not-quite-pure-yellow is:        255,   0, 0 to 255, 254,   0
+  // Pure yellow to not-quite-pure-green is: 255, 255, 0 to   1, 255,   0
+  // Pure green to not-quite-pure-cyan is:     0, 255, 0 to   0, 255, 254
+  // and so forth. Hence, 1530 distinct hues (0 to 1529), and hence why
+  // the constants below are not the multiples of 256 you might expect.
+
+  // Convert hue to R,G,B (nested ifs faster than divide+mod+switch):
+  if(hue < 510) {         // Red to Green-1
+    b = 0;
+    if(hue < 255) {       //   Red to Yellow-1
+      r = 255;
+      g = hue;            //     g = 0 to 254
+    } else {              //   Yellow to Green-1
+      r = 510 - hue;      //     r = 255 to 1
+      g = 255;
+    }
+  } else if(hue < 1020) { // Green to Blue-1
+    r = 0;
+    if(hue <  765) {      //   Green to Cyan-1
+      g = 255;
+      b = hue - 510;      //     b = 0 to 254
+    } else {              //   Cyan to Blue-1
+      g = 1020 - hue;     //     g = 255 to 1
+      b = 255;
+    }
+  } else if(hue < 1530) { // Blue to Red-1
+    g = 0;
+    if(hue < 1275) {      //   Blue to Magenta-1
+      r = hue - 1020;     //     r = 0 to 254
+      b = 255;
+    } else {              //   Magenta to Red-1
+      r = 255;
+      b = 1530 - hue;     //     b = 255 to 1
+    }
+  } else {                // Last 0.5 Red (quicker than % operator)
+    r = 255;
+    g = b = 0;
+  }
+
+  // Apply saturation and value to R,G,B, pack into 32-bit result:
+  uint32_t v1 =   1 + val; // 1 to 256; allows >>8 instead of /255
+  uint16_t s1 =   1 + sat; // 1 to 256; same reason
+  uint8_t  s2 = 255 - sat; // 255 to 0
+  return ((((((r * s1) >> 8) + s2) * v1) & 0xff00) << 8) |
+          (((((g * s1) >> 8) + s2) * v1) & 0xff00)       |
+         ( ((((b * s1) >> 8) + s2) * v1)           >> 8);
+}
+
+/*!
+  @brief   Query the color of a previously-set pixel.
+  @param   n  Index of pixel to read (0 = first).
+  @return  'Packed' 32-bit RGB or WRGB value. Most significant byte is white
+           (for RGBW pixels) or 0 (for RGB pixels), next is red, then green,
+           and least significant byte is blue.
+  @note    If the strip brightness has been changed from the default value
+           of 255, the color read from a pixel may not exactly match what
+           was previously written with one of the setPixelColor() functions.
+           This gets more pronounced at lower brightness levels.
+*/
+uint32_t Adafruit_NeoPixel::getPixelColor(uint16_t n) const {
+  if(n >= numLEDs) return 0; // Out of bounds, return no color.
+
+  uint8_t *p;
+
+  if(wOffset == rOffset) { // Is RGB-type device
+    p = &pixels[n * 3];
+    if(brightness) {
+      // Stored color was decimated by setBrightness(). Returned value
+      // attempts to scale back to an approximation of the original 24-bit
+      // value used when setting the pixel color, but there will always be
+      // some error -- those bits are simply gone. Issue is most
+      // pronounced at low brightness levels.
+      return (((uint32_t)(p[rOffset] << 8) / brightness) << 16) |
+             (((uint32_t)(p[gOffset] << 8) / brightness) <<  8) |
+             ( (uint32_t)(p[bOffset] << 8) / brightness       );
+    } else {
+      // No brightness adjustment has been made -- return 'raw' color
+      return ((uint32_t)p[rOffset] << 16) |
+             ((uint32_t)p[gOffset] <<  8) |
+              (uint32_t)p[bOffset];
+    }
+  } else {                 // Is RGBW-type device
+    p = &pixels[n * 4];
+    if(brightness) { // Return scaled color
+      return (((uint32_t)(p[wOffset] << 8) / brightness) << 24) |
+             (((uint32_t)(p[rOffset] << 8) / brightness) << 16) |
+             (((uint32_t)(p[gOffset] << 8) / brightness) <<  8) |
+             ( (uint32_t)(p[bOffset] << 8) / brightness       );
+    } else { // Return raw color
+      return ((uint32_t)p[wOffset] << 24) |
+             ((uint32_t)p[rOffset] << 16) |
+             ((uint32_t)p[gOffset] <<  8) |
+              (uint32_t)p[bOffset];
+    }
+  }
+}
+
+
+/*!
+  @brief   Adjust output brightness. Does not immediately affect what's
+           currently displayed on the LEDs. The next call to show() will
+           refresh the LEDs at this level.
+  @param   b  Brightness setting, 0=minimum (off), 255=brightest.
+  @note    This was intended for one-time use in one's setup() function,
+           not as an animation effect in itself. Because of the way this
+           library "pre-multiplies" LED colors in RAM, changing the
+           brightness is often a "lossy" operation -- what you write to
+           pixels isn't necessary the same as what you'll read back.
+           Repeated brightness changes using this function exacerbate the
+           problem. Smart programs therefore treat the strip as a
+           write-only resource, maintaining their own state to render each
+           frame of an animation, not relying on read-modify-write.
+*/
+void Adafruit_NeoPixel::setBrightness(uint8_t b) {
+  // Stored brightness value is different than what's passed.
+  // This simplifies the actual scaling math later, allowing a fast
+  // 8x8-bit multiply and taking the MSB. 'brightness' is a uint8_t,
+  // adding 1 here may (intentionally) roll over...so 0 = max brightness
+  // (color values are interpreted literally; no scaling), 1 = min
+  // brightness (off), 255 = just below max brightness.
+  uint8_t newBrightness = b + 1;
+  if(newBrightness != brightness) { // Compare against prior value
+    // Brightness has changed -- re-scale existing data in RAM,
+    // This process is potentially "lossy," especially when increasing
+    // brightness. The tight timing in the WS2811/WS2812 code means there
+    // aren't enough free cycles to perform this scaling on the fly as data
+    // is issued. So we make a pass through the existing color data in RAM
+    // and scale it (subsequent graphics commands also work at this
+    // brightness level). If there's a significant step up in brightness,
+    // the limited number of steps (quantization) in the old data will be
+    // quite visible in the re-scaled version. For a non-destructive
+    // change, you'll need to re-render the full strip data. C'est la vie.
+    uint8_t  c,
+            *ptr           = pixels,
+             oldBrightness = brightness - 1; // De-wrap old brightness value
+    uint16_t scale;
+    if(oldBrightness == 0) scale = 0; // Avoid /0
+    else if(b == 255) scale = 65535 / oldBrightness;
+    else scale = (((uint16_t)newBrightness << 8) - 1) / oldBrightness;
+    for(uint16_t i=0; i<numBytes; i++) {
+      c      = *ptr;
+      *ptr++ = (c * scale) >> 8;
+    }
+    brightness = newBrightness;
+  }
+}
+
+/*!
+  @brief   Retrieve the last-set brightness value for the strip.
+  @return  Brightness value: 0 = minimum (off), 255 = maximum.
+*/
+uint8_t Adafruit_NeoPixel::getBrightness(void) const {
+  return brightness - 1;
+}
+
+/*!
+  @brief   Fill the whole NeoPixel strip with 0 / black / off.
+*/
+void Adafruit_NeoPixel::clear(void) {
+  memset(pixels, 0, numBytes);
+}
+
+// A 32-bit variant of gamma8() that applies the same function
+// to all components of a packed RGB or WRGB value.
+uint32_t Adafruit_NeoPixel::gamma32(uint32_t x) {
+  uint8_t *y = (uint8_t *)&x;
+  // All four bytes of a 32-bit value are filtered even if RGB (not WRGB),
+  // to avoid a bunch of shifting and masking that would be necessary for
+  // properly handling different endianisms (and each byte is a fairly
+  // trivial operation, so it might not even be wasting cycles vs a check
+  // and branch for the RGB case). In theory this might cause trouble *if*
+  // someone's storing information in the unused most significant byte
+  // of an RGB value, but this seems exceedingly rare and if it's
+  // encountered in reality they can mask values going in or coming out.
+  for(uint8_t i=0; i<4; i++) y[i] = gamma8(y[i]);
+  return x; // Packed 32-bit return
+}

Adafruit_NeoPixel/Adafruit_NeoPixel.h

@@ -0,0 +1,386 @@
+/*!
+ * @file Adafruit_NeoPixel.h
+ *
+ * This is part of Adafruit's NeoPixel library for the Arduino platform,
+ * allowing a broad range of microcontroller boards (most AVR boards,
+ * many ARM devices, ESP8266 and ESP32, among others) to control Adafruit
+ * NeoPixels, FLORA RGB Smart Pixels and compatible devices -- WS2811,
+ * WS2812, WS2812B, SK6812, etc.
+ *
+ * Adafruit invests time and resources providing this open source code,
+ * please support Adafruit and open-source hardware by purchasing products
+ * from Adafruit!
+ *
+ * Written by Phil "Paint Your Dragon" Burgess for Adafruit Industries,
+ * with contributions by PJRC, Michael Miller and other members of the
+ * open source community.
+ *
+ * This file is part of the Adafruit_NeoPixel library.
+ *
+ * Adafruit_NeoPixel is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public License as
+ * published by the Free Software Foundation, either version 3 of the
+ * License, or (at your option) any later version.
+ *
+ * Adafruit_NeoPixel is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with NeoPixel.  If not, see
+ * <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#ifndef ADAFRUIT_NEOPIXEL_H
+#define ADAFRUIT_NEOPIXEL_H
+
+#ifdef ARDUINO
+  #if (ARDUINO >= 100)
+  #include <Arduino.h>
+  #else
+  #include <WProgram.h>
+  #include <pins_arduino.h>
+  #endif
+#endif
+
+#ifdef TARGET_LPC1768
+  #include <Arduino.h>
+#endif
+
+// The order of primary colors in the NeoPixel data stream can vary among
+// device types, manufacturers and even different revisions of the same
+// item.  The third parameter to the Adafruit_NeoPixel constructor encodes
+// the per-pixel byte offsets of the red, green and blue primaries (plus
+// white, if present) in the data stream -- the following #defines provide
+// an easier-to-use named version for each permutation. e.g. NEO_GRB
+// indicates a NeoPixel-compatible device expecting three bytes per pixel,
+// with the first byte transmitted containing the green value, second
+// containing red and third containing blue. The in-memory representation
+// of a chain of NeoPixels is the same as the data-stream order; no
+// re-ordering of bytes is required when issuing data to the chain.
+// Most of these values won't exist in real-world devices, but it's done
+// this way so we're ready for it (also, if using the WS2811 driver IC,
+// one might have their pixels set up in any weird permutation).
+
+// Bits 5,4 of this value are the offset (0-3) from the first byte of a
+// pixel to the location of the red color byte.  Bits 3,2 are the green
+// offset and 1,0 are the blue offset.  If it is an RGBW-type device
+// (supporting a white primary in addition to R,G,B), bits 7,6 are the
+// offset to the white byte...otherwise, bits 7,6 are set to the same value
+// as 5,4 (red) to indicate an RGB (not RGBW) device.
+// i.e. binary representation:
+// 0bWWRRGGBB for RGBW devices
+// 0bRRRRGGBB for RGB
+
+// RGB NeoPixel permutations; white and red offsets are always same
+// Offset:         W        R        G        B
+#define NEO_RGB  ((0<<6) | (0<<4) | (1<<2) | (2)) ///< Transmit as R,G,B
+#define NEO_RBG  ((0<<6) | (0<<4) | (2<<2) | (1)) ///< Transmit as R,B,G
+#define NEO_GRB  ((1<<6) | (1<<4) | (0<<2) | (2)) ///< Transmit as G,R,B
+#define NEO_GBR  ((2<<6) | (2<<4) | (0<<2) | (1)) ///< Transmit as G,B,R
+#define NEO_BRG  ((1<<6) | (1<<4) | (2<<2) | (0)) ///< Transmit as B,R,G
+#define NEO_BGR  ((2<<6) | (2<<4) | (1<<2) | (0)) ///< Transmit as B,G,R
+
+// RGBW NeoPixel permutations; all 4 offsets are distinct
+// Offset:         W          R          G          B
+#define NEO_WRGB ((0<<6) | (1<<4) | (2<<2) | (3)) ///< Transmit as W,R,G,B
+#define NEO_WRBG ((0<<6) | (1<<4) | (3<<2) | (2)) ///< Transmit as W,R,B,G
+#define NEO_WGRB ((0<<6) | (2<<4) | (1<<2) | (3)) ///< Transmit as W,G,R,B
+#define NEO_WGBR ((0<<6) | (3<<4) | (1<<2) | (2)) ///< Transmit as W,G,B,R
+#define NEO_WBRG ((0<<6) | (2<<4) | (3<<2) | (1)) ///< Transmit as W,B,R,G
+#define NEO_WBGR ((0<<6) | (3<<4) | (2<<2) | (1)) ///< Transmit as W,B,G,R
+
+#define NEO_RWGB ((1<<6) | (0<<4) | (2<<2) | (3)) ///< Transmit as R,W,G,B
+#define NEO_RWBG ((1<<6) | (0<<4) | (3<<2) | (2)) ///< Transmit as R,W,B,G
+#define NEO_RGWB ((2<<6) | (0<<4) | (1<<2) | (3)) ///< Transmit as R,G,W,B
+#define NEO_RGBW ((3<<6) | (0<<4) | (1<<2) | (2)) ///< Transmit as R,G,B,W
+#define NEO_RBWG ((2<<6) | (0<<4) | (3<<2) | (1)) ///< Transmit as R,B,W,G
+#define NEO_RBGW ((3<<6) | (0<<4) | (2<<2) | (1)) ///< Transmit as R,B,G,W
+
+#define NEO_GWRB ((1<<6) | (2<<4) | (0<<2) | (3)) ///< Transmit as G,W,R,B
+#define NEO_GWBR ((1<<6) | (3<<4) | (0<<2) | (2)) ///< Transmit as G,W,B,R
+#define NEO_GRWB ((2<<6) | (1<<4) | (0<<2) | (3)) ///< Transmit as G,R,W,B
+#define NEO_GRBW ((3<<6) | (1<<4) | (0<<2) | (2)) ///< Transmit as G,R,B,W
+#define NEO_GBWR ((2<<6) | (3<<4) | (0<<2) | (1)) ///< Transmit as G,B,W,R
+#define NEO_GBRW ((3<<6) | (2<<4) | (0<<2) | (1)) ///< Transmit as G,B,R,W
+
+#define NEO_BWRG ((1<<6) | (2<<4) | (3<<2) | (0)) ///< Transmit as B,W,R,G
+#define NEO_BWGR ((1<<6) | (3<<4) | (2<<2) | (0)) ///< Transmit as B,W,G,R
+#define NEO_BRWG ((2<<6) | (1<<4) | (3<<2) | (0)) ///< Transmit as B,R,W,G
+#define NEO_BRGW ((3<<6) | (1<<4) | (2<<2) | (0)) ///< Transmit as B,R,G,W
+#define NEO_BGWR ((2<<6) | (3<<4) | (1<<2) | (0)) ///< Transmit as B,G,W,R
+#define NEO_BGRW ((3<<6) | (2<<4) | (1<<2) | (0)) ///< Transmit as B,G,R,W
+
+// Add NEO_KHZ400 to the color order value to indicate a 400 KHz device.
+// All but the earliest v1 NeoPixels expect an 800 KHz data stream, this is
+// the default if unspecified. Because flash space is very limited on ATtiny
+// devices (e.g. Trinket, Gemma), v1 NeoPixels aren't handled by default on
+// those chips, though it can be enabled by removing the ifndef/endif below,
+// but code will be bigger. Conversely, can disable the NEO_KHZ400 line on
+// other MCUs to remove v1 support and save a little space.
+
+#define NEO_KHZ800 0x0000 ///< 800 KHz data transmission
+#ifndef __AVR_ATtiny85__
+#define NEO_KHZ400 0x0100 ///< 400 KHz data transmission
+#endif
+
+// If 400 KHz support is enabled, the third parameter to the constructor
+// requires a 16-bit value (in order to select 400 vs 800 KHz speed).
+// If only 800 KHz is enabled (as is default on ATtiny), an 8-bit value
+// is sufficient to encode pixel color order, saving some space.
+
+#ifdef NEO_KHZ400
+typedef uint16_t neoPixelType; ///< 3rd arg to Adafruit_NeoPixel constructor
+#else
+typedef uint8_t  neoPixelType; ///< 3rd arg to Adafruit_NeoPixel constructor
+#endif
+
+// These two tables are declared outside the Adafruit_NeoPixel class
+// because some boards may require oldschool compilers that don't
+// handle the C++11 constexpr keyword.
+
+/* A PROGMEM (flash mem) table containing 8-bit unsigned sine wave (0-255).
+   Copy & paste this snippet into a Python REPL to regenerate:
+import math
+for x in range(256):
+    print("{:3},".format(int((math.sin(x/128.0*math.pi)+1.0)*127.5+0.5))),
+    if x&15 == 15: print
+*/
+/*
+static const uint8_t PROGMEM _NeoPixelSineTable[256] = {
+  128,131,134,137,140,143,146,149,152,155,158,162,165,167,170,173,
+  176,179,182,185,188,190,193,196,198,201,203,206,208,211,213,215,
+  218,220,222,224,226,228,230,232,234,235,237,238,240,241,243,244,
+  245,246,248,249,250,250,251,252,253,253,254,254,254,255,255,255,
+  255,255,255,255,254,254,254,253,253,252,251,250,250,249,248,246,
+  245,244,243,241,240,238,237,235,234,232,230,228,226,224,222,220,
+  218,215,213,211,208,206,203,201,198,196,193,190,188,185,182,179,
+  176,173,170,167,165,162,158,155,152,149,146,143,140,137,134,131,
+  128,124,121,118,115,112,109,106,103,100, 97, 93, 90, 88, 85, 82,
+   79, 76, 73, 70, 67, 65, 62, 59, 57, 54, 52, 49, 47, 44, 42, 40,
+   37, 35, 33, 31, 29, 27, 25, 23, 21, 20, 18, 17, 15, 14, 12, 11,
+   10,  9,  7,  6,  5,  5,  4,  3,  2,  2,  1,  1,  1,  0,  0,  0,
+    0,  0,  0,  0,  1,  1,  1,  2,  2,  3,  4,  5,  5,  6,  7,  9,
+   10, 11, 12, 14, 15, 17, 18, 20, 21, 23, 25, 27, 29, 31, 33, 35,
+   37, 40, 42, 44, 47, 49, 52, 54, 57, 59, 62, 65, 67, 70, 73, 76,
+   79, 82, 85, 88, 90, 93, 97,100,103,106,109,112,115,118,121,124};
+
+*/
+/* Similar to above, but for an 8-bit gamma-correction table.
+   Copy & paste this snippet into a Python REPL to regenerate:
+import math
+gamma=2.6
+for x in range(256):
+    print("{:3},".format(int(math.pow((x)/255.0,gamma)*255.0+0.5))),
+    if x&15 == 15: print
+*//*
+
+static const uint8_t PROGMEM _NeoPixelGammaTable[256] = {
+    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
+    0,  0,  0,  0,  0,  0,  0,  0,  1,  1,  1,  1,  1,  1,  1,  1,
+    1,  1,  1,  1,  2,  2,  2,  2,  2,  2,  2,  2,  3,  3,  3,  3,
+    3,  3,  4,  4,  4,  4,  5,  5,  5,  5,  5,  6,  6,  6,  6,  7,
+    7,  7,  8,  8,  8,  9,  9,  9, 10, 10, 10, 11, 11, 11, 12, 12,
+   13, 13, 13, 14, 14, 15, 15, 16, 16, 17, 17, 18, 18, 19, 19, 20,
+   20, 21, 21, 22, 22, 23, 24, 24, 25, 25, 26, 27, 27, 28, 29, 29,
+   30, 31, 31, 32, 33, 34, 34, 35, 36, 37, 38, 38, 39, 40, 41, 42,
+   42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57,
+   58, 59, 60, 61, 62, 63, 64, 65, 66, 68, 69, 70, 71, 72, 73, 75,
+   76, 77, 78, 80, 81, 82, 84, 85, 86, 88, 89, 90, 92, 93, 94, 96,
+   97, 99,100,102,103,105,106,108,109,111,112,114,115,117,119,120,
+  122,124,125,127,129,130,132,134,136,137,139,141,143,145,146,148,
+  150,152,154,156,158,160,162,164,166,168,170,172,174,176,178,180,
+  182,184,186,188,191,193,195,197,199,202,204,206,209,211,213,215,
+  218,220,223,225,227,230,232,235,237,240,242,245,247,250,252,255};
+*/
+
+/*! 
+    @brief  Class that stores state and functions for interacting with
+            Adafruit NeoPixels and compatible devices.
+*/
+class Adafruit_NeoPixel {
+
+ public:
+
+  // Constructor: number of LEDs, pin number, LED type
+  Adafruit_NeoPixel(uint16_t n, int16_t pin=6,
+    neoPixelType type=NEO_GRB + NEO_KHZ800);
+  Adafruit_NeoPixel(void);
+  ~Adafruit_NeoPixel();
+
+  void              begin(void);
+  void              show(void);
+  void              setPin(int16_t p);
+  void              setPixelColor(uint16_t n, uint8_t r, uint8_t g, uint8_t b);
+  void              setPixelColor(uint16_t n, uint8_t r, uint8_t g, uint8_t b,
+                      uint8_t w);
+  void              setPixelColor(uint16_t n, uint32_t c);
+  void              fill(uint32_t c=0, uint16_t first=0, uint16_t count=0);
+  void              setBrightness(uint8_t);
+  void              clear(void);
+  void              updateLength(uint16_t n);
+  void              updateType(neoPixelType t);
+  /*!
+    @brief   Check whether a call to show() will start sending data
+             immediately or will 'block' for a required interval. NeoPixels
+             require a short quiet time (about 300 microseconds) after the
+             last bit is received before the data 'latches' and new data can
+             start being received. Usually one's sketch is implicitly using
+             this time to generate a new frame of animation...but if it
+             finishes very quickly, this function could be used to see if
+             there's some idle time available for some low-priority
+             concurrent task.
+    @return  1 or true if show() will start sending immediately, 0 or false
+             if show() would block (meaning some idle time is available).
+  */
+  bool canShow(void) {
+    // It's normal and possible for endTime to exceed micros() if the
+    // 32-bit clock counter has rolled over (about every 70 minutes).
+    // Since both are uint32_t, a negative delta correctly maps back to
+    // positive space, and it would seem like the subtraction below would
+    // suffice. But a problem arises if code invokes show() very
+    // infrequently...the micros() counter may roll over MULTIPLE times in
+    // that interval, the delta calculation is no longer correct and the
+    // next update may stall for a very long time. The check below resets
+    // the latch counter if a rollover has occurred. This can cause an
+    // extra delay of up to 300 microseconds in the rare case where a
+    // show() call happens precisely around the rollover, but that's
+    // neither likely nor especially harmful, vs. other code that might
+    // stall for 30+ minutes, or having to document and frequently remind
+    // and/or provide tech support explaining an unintuitive need for
+    // show() calls at least once an hour.
+    uint32_t now = micros();
+    if (endTime > now) {
+      endTime = now;
+    }
+    return (now - endTime) >= 300L;
+  }
+  /*!
+    @brief   Get a pointer directly to the NeoPixel data buffer in RAM.
+             Pixel data is stored in a device-native format (a la the NEO_*
+             constants) and is not translated here. Applications that access
+             this buffer will need to be aware of the specific data format
+             and handle colors appropriately.
+    @return  Pointer to NeoPixel buffer (uint8_t* array).
+    @note    This is for high-performance applications where calling
+             setPixelColor() on every single pixel would be too slow (e.g.
+             POV or light-painting projects). There is no bounds checking
+             on the array, creating tremendous potential for mayhem if one
+             writes past the ends of the buffer. Great power, great
+             responsibility and all that.
+  */
+  uint8_t          *getPixels(void) const { return pixels; };
+  uint8_t           getBrightness(void) const;
+  /*!
+    @brief   Retrieve the pin number used for NeoPixel data output.
+    @return  Arduino pin number (-1 if not set).
+  */
+  int16_t           getPin(void) const { return pin; };
+  /*!
+    @brief   Return the number of pixels in an Adafruit_NeoPixel strip object.
+    @return  Pixel count (0 if not set).
+  */
+  uint16_t          numPixels(void) const { return numLEDs; }
+  uint32_t          getPixelColor(uint16_t n) const;
+  /*!
+    @brief   An 8-bit integer sine wave function, not directly compatible
+             with standard trigonometric units like radians or degrees.
+    @param   x  Input angle, 0-255; 256 would loop back to zero, completing
+                the circle (equivalent to 360 degrees or 2 pi radians).
+                One can therefore use an unsigned 8-bit variable and simply
+                add or subtract, allowing it to overflow/underflow and it
+                still does the expected contiguous thing.
+    @return  Sine result, 0 to 255, or -128 to +127 if type-converted to
+             a signed int8_t, but you'll most likely want unsigned as this
+             output is often used for pixel brightness in animation effects.
+  */
+  static uint8_t    sine8(uint8_t x) {
+    return nullptr; // 0-255 in, 0-255 out
+  }
+  /*!
+    @brief   An 8-bit gamma-correction function for basic pixel brightness
+             adjustment. Makes color transitions appear more perceptially
+             correct.
+    @param   x  Input brightness, 0 (minimum or off/black) to 255 (maximum).
+    @return  Gamma-adjusted brightness, can then be passed to one of the
+             setPixelColor() functions. This uses a fixed gamma correction
+             exponent of 2.6, which seems reasonably okay for average
+             NeoPixels in average tasks. If you need finer control you'll
+             need to provide your own gamma-correction function instead.
+  */
+  static uint8_t    gamma8(uint8_t x) {
+    return nullptr; // 0-255 in, 0-255 out
+  }
+  /*!
+    @brief   Convert separate red, green and blue values into a single
+             "packed" 32-bit RGB color.
+    @param   r  Red brightness, 0 to 255.
+    @param   g  Green brightness, 0 to 255.
+    @param   b  Blue brightness, 0 to 255.
+    @return  32-bit packed RGB value, which can then be assigned to a
+             variable for later use or passed to the setPixelColor()
+             function. Packed RGB format is predictable, regardless of
+             LED strand color order.
+  */
+  static uint32_t   Color(uint8_t r, uint8_t g, uint8_t b) {
+    return ((uint32_t)r << 16) | ((uint32_t)g <<  8) | b;
+  }
+  /*!
+    @brief   Convert separate red, green, blue and white values into a
+             single "packed" 32-bit WRGB color.
+    @param   r  Red brightness, 0 to 255.
+    @param   g  Green brightness, 0 to 255.
+    @param   b  Blue brightness, 0 to 255.
+    @param   w  White brightness, 0 to 255.
+    @return  32-bit packed WRGB value, which can then be assigned to a
+             variable for later use or passed to the setPixelColor()
+             function. Packed WRGB format is predictable, regardless of
+             LED strand color order.
+  */
+  static uint32_t   Color(uint8_t r, uint8_t g, uint8_t b, uint8_t w) {
+    return ((uint32_t)w << 24) | ((uint32_t)r << 16) | ((uint32_t)g <<  8) | b;
+  }
+  static uint32_t   ColorHSV(uint16_t hue, uint8_t sat=255, uint8_t val=255);
+  /*!
+    @brief   A gamma-correction function for 32-bit packed RGB or WRGB
+             colors. Makes color transitions appear more perceptially
+             correct.
+    @param   x  32-bit packed RGB or WRGB color.
+    @return  Gamma-adjusted packed color, can then be passed in one of the
+             setPixelColor() functions. Like gamma8(), this uses a fixed
+             gamma correction exponent of 2.6, which seems reasonably okay
+             for average NeoPixels in average tasks. If you need finer
+             control you'll need to provide your own gamma-correction
+             function instead.
+  */
+  static uint32_t   gamma32(uint32_t x);
+
+ protected:
+
+#ifdef NEO_KHZ400  // If 400 KHz NeoPixel support enabled...
+  bool              is800KHz;   ///< true if 800 KHz pixels
+#endif
+  bool              begun;      ///< true if begin() previously called
+  uint16_t          numLEDs;    ///< Number of RGB LEDs in strip
+  uint16_t          numBytes;   ///< Size of 'pixels' buffer below
+  int16_t           pin;        ///< Output pin number (-1 if not yet set)
+  uint8_t           brightness; ///< Strip brightness 0-255 (stored as +1)
+  uint8_t          *pixels;     ///< Holds LED color values (3 or 4 bytes each)
+  uint8_t           rOffset;    ///< Red index within each 3- or 4-byte pixel
+  uint8_t           gOffset;    ///< Index of green byte
+  uint8_t           bOffset;    ///< Index of blue byte
+  uint8_t           wOffset;    ///< Index of white (==rOffset if no white)
+  uint32_t          endTime;    ///< Latch timing reference
+#ifdef __AVR__
+  volatile uint8_t *port;       ///< Output PORT register
+  uint8_t           pinMask;    ///< Output PORT bitmask
+#endif
+#if defined(ARDUINO_ARCH_STM32) || defined(ARDUINO_ARCH_ARDUINO_CORE_STM32)
+  GPIO_TypeDef *gpioPort;       ///< Output GPIO PORT
+  uint32_t gpioPin;             ///< Output GPIO PIN
+#endif
+};
+
+#endif // ADAFRUIT_NEOPIXEL_H

Adafruit_NeoPixel/CONTRIBUTING.md

@@ -0,0 +1,13 @@
+# Contribution Guidelines
+
+This library is the culmination of the expertise of many members of the open source community who have dedicated their time and hard work. The best way to ask for help or propose a new idea is to [create a new issue](https://github.com/adafruit/Adafruit_NeoPixel/issues/new) while creating a Pull Request with your code changes allows you to share your own innovations with the rest of the community.
+
+The following are some guidelines to observe when creating issues or PRs:
+
+- Be friendly; it is important that we can all enjoy a safe space as we are all working on the same project and it is okay for people to have different ideas
+
+- [Use code blocks](https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet#code); it helps us help you when we can read your code! On that note also refrain from pasting more than 30 lines of code in a post, instead [create a gist](https://gist.github.com/) if you need to share large snippets
+
+- Use reasonable titles; refrain from using overly long or capitalized titles as they are usually annoying and do little to encourage others to help :smile:
+
+- Be detailed; refrain from mentioning code problems without sharing your source code and always give information regarding your board and version of the library

Adafruit_NeoPixel/COPYING

@@ -0,0 +1,165 @@
+                   GNU LESSER GENERAL PUBLIC LICENSE
+                       Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+
+  This version of the GNU Lesser General Public License incorporates
+the terms and conditions of version 3 of the GNU General Public
+License, supplemented by the additional permissions listed below.
+
+  0. Additional Definitions.
+
+  As used herein, "this License" refers to version 3 of the GNU Lesser
+General Public License, and the "GNU GPL" refers to version 3 of the GNU
+General Public License.
+
+  "The Library" refers to a covered work governed by this License,
+other than an Application or a Combined Work as defined below.
+
+  An "Application" is any work that makes use of an interface provided
+by the Library, but which is not otherwise based on the Library.
+Defining a subclass of a class defined by the Library is deemed a mode
+of using an interface provided by the Library.
+
+  A "Combined Work" is a work produced by combining or linking an
+Application with the Library.  The particular version of the Library
+with which the Combined Work was made is also called the "Linked
+Version".
+
+  The "Minimal Corresponding Source" for a Combined Work means the
+Corresponding Source for the Combined Work, excluding any source code
+for portions of the Combined Work that, considered in isolation, are
+based on the Application, and not on the Linked Version.
+
+  The "Corresponding Application Code" for a Combined Work means the
+object code and/or source code for the Application, including any data
+and utility programs needed for reproducing the Combined Work from the
+Application, but excluding the System Libraries of the Combined Work.
+
+  1. Exception to Section 3 of the GNU GPL.
+
+  You may convey a covered work under sections 3 and 4 of this License
+without being bound by section 3 of the GNU GPL.
+
+  2. Conveying Modified Versions.
+
+  If you modify a copy of the Library, and, in your modifications, a
+facility refers to a function or data to be supplied by an Application
+that uses the facility (other than as an argument passed when the
+facility is invoked), then you may convey a copy of the modified
+version:
+
+   a) under this License, provided that you make a good faith effort to
+   ensure that, in the event an Application does not supply the
+   function or data, the facility still operates, and performs
+   whatever part of its purpose remains meaningful, or
+
+   b) under the GNU GPL, with none of the additional permissions of
+   this License applicable to that copy.
+
+  3. Object Code Incorporating Material from Library Header Files.
+
+  The object code form of an Application may incorporate material from
+a header file that is part of the Library.  You may convey such object
+code under terms of your choice, provided that, if the incorporated
+material is not limited to numerical parameters, data structure
+layouts and accessors, or small macros, inline functions and templates
+(ten or fewer lines in length), you do both of the following:
+
+   a) Give prominent notice with each copy of the object code that the
+   Library is used in it and that the Library and its use are
+   covered by this License.
+
+   b) Accompany the object code with a copy of the GNU GPL and this license
+   document.
+
+  4. Combined Works.
+
+  You may convey a Combined Work under terms of your choice that,
+taken together, effectively do not restrict modification of the
+portions of the Library contained in the Combined Work and reverse
+engineering for debugging such modifications, if you also do each of
+the following:
+
+   a) Give prominent notice with each copy of the Combined Work that
+   the Library is used in it and that the Library and its use are
+   covered by this License.
+
+   b) Accompany the Combined Work with a copy of the GNU GPL and this license
+   document.
+
+   c) For a Combined Work that displays copyright notices during
+   execution, include the copyright notice for the Library among
+   these notices, as well as a reference directing the user to the
+   copies of the GNU GPL and this license document.
+
+   d) Do one of the following:
+
+       0) Convey the Minimal Corresponding Source under the terms of this
+       License, and the Corresponding Application Code in a form
+       suitable for, and under terms that permit, the user to
+       recombine or relink the Application with a modified version of
+       the Linked Version to produce a modified Combined Work, in the
+       manner specified by section 6 of the GNU GPL for conveying
+       Corresponding Source.
+
+       1) Use a suitable shared library mechanism for linking with the
+       Library.  A suitable mechanism is one that (a) uses at run time
+       a copy of the Library already present on the user's computer
+       system, and (b) will operate properly with a modified version
+       of the Library that is interface-compatible with the Linked
+       Version.
+
+   e) Provide Installation Information, but only if you would otherwise
+   be required to provide such information under section 6 of the
+   GNU GPL, and only to the extent that such information is
+   necessary to install and execute a modified version of the
+   Combined Work produced by recombining or relinking the
+   Application with a modified version of the Linked Version. (If
+   you use option 4d0, the Installation Information must accompany
+   the Minimal Corresponding Source and Corresponding Application
+   Code. If you use option 4d1, you must provide the Installation
+   Information in the manner specified by section 6 of the GNU GPL
+   for conveying Corresponding Source.)
+
+  5. Combined Libraries.
+
+  You may place library facilities that are a work based on the
+Library side by side in a single library together with other library
+facilities that are not Applications and are not covered by this
+License, and convey such a combined library under terms of your
+choice, if you do both of the following:
+
+   a) Accompany the combined library with a copy of the same work based
+   on the Library, uncombined with any other library facilities,
+   conveyed under the terms of this License.
+
+   b) Give prominent notice with the combined library that part of it
+   is a work based on the Library, and explaining where to find the
+   accompanying uncombined form of the same work.
+
+  6. Revised Versions of the GNU Lesser General Public License.
+
+  The Free Software Foundation may publish revised and/or new versions
+of the GNU Lesser General Public License from time to time. Such new
+versions will be similar in spirit to the present version, but may
+differ in detail to address new problems or concerns.
+
+  Each version is given a distinguishing version number. If the
+Library as you received it specifies that a certain numbered version
+of the GNU Lesser General Public License "or any later version"
+applies to it, you have the option of following the terms and
+conditions either of that published version or of any later version
+published by the Free Software Foundation. If the Library as you
+received it does not specify a version number of the GNU Lesser
+General Public License, you may choose any version of the GNU Lesser
+General Public License ever published by the Free Software Foundation.
+
+  If the Library as you received it specifies that a proxy can decide
+whether future versions of the GNU Lesser General Public License shall
+apply, that proxy's public statement of acceptance of any version is
+permanent authorization for you to choose that version for the
+Library.

Adafruit_NeoPixel/README.md

@@ -0,0 +1,153 @@
+# Adafruit NeoPixel Library [![Build Status](https://github.com/adafruit/Adafruit_NeoPixel/workflows/Arduino%20Library%20CI/badge.svg)](https://github.com/adafruit/Adafruit_NeoPixel/actions)[![Documentation](https://github.com/adafruit/ci-arduino/blob/master/assets/doxygen_badge.svg)](http://adafruit.github.io/Adafruit_NeoPixel/html/index.html)
+
+Arduino library for controlling single-wire-based LED pixels and strip such as the [Adafruit 60 LED/meter Digital LED strip][strip], the [Adafruit FLORA RGB Smart Pixel][flora], the [Adafruit Breadboard-friendly RGB Smart Pixel][pixel], the [Adafruit NeoPixel Stick][stick], and the [Adafruit NeoPixel Shield][shield].
+
+After downloading, rename folder to 'Adafruit_NeoPixel' and install in Arduino Libraries folder. Restart Arduino IDE, then open File->Sketchbook->Library->Adafruit_NeoPixel->strandtest sketch.
+
+Compatibility notes: Port A is not supported on any AVR processors at this time
+
+[flora]:  http://adafruit.com/products/1060
+[strip]:  http://adafruit.com/products/1138
+[pixel]:  http://adafruit.com/products/1312
+[stick]:  http://adafruit.com/products/1426
+[shield]: http://adafruit.com/products/1430
+
+---
+
+## Installation
+
+### First Method
+
+![image](https://user-images.githubusercontent.com/36513474/68967967-3e37f480-0803-11ea-91d9-601848c306ee.png)
+
+1. In the Arduino IDE, navigate to Sketch > Include Library > Manage Libraries
+1. Then the Library Manager will open and you will find a list of libraries that are already installed or ready for installation.
+1. Then search for Neopixel strip using the search bar.
+1. Click on the text area and then select the specific version and install it.
+
+### Second Method
+
+1. Navigate to the [Releases page](https://github.com/adafruit/Adafruit_NeoPixel/releases).
+1. Download the latest release.
+1. Extract the zip file
+1. In the Arduino IDE, navigate to Sketch > Include Library > Add .ZIP Library
+
+## Features
+
+- ### Simple to use
+
+  Controlling NeoPixels “from scratch” is quite a challenge, so we provide a library letting you focus on the fun and interesting bits.
+
+- ### Give back
+
+  The library is free; you don’t have to pay for anything. Adafruit invests time and resources providing this open source code, please support Adafruit and open-source hardware by purchasing products from Adafruit!
+
+- ### Supported Chipsets
+
+  We have included code for the following chips - sometimes these break for exciting reasons that we can't control in which case please open an issue!
+
+  - AVR ATmega and ATtiny (any 8-bit) - 8 MHz, 12 MHz and 16 MHz
+  - Teensy 3.x and LC
+  - Arduino Due
+  - Arduino 101
+  - ATSAMD21 (Arduino Zero/M0 and other SAMD21 boards) @ 48 MHz
+  - ATSAMD51 @ 120 MHz
+  - Adafruit STM32 Feather @ 120 MHz
+  - ESP8266 any speed
+  - ESP32 any speed
+  - Nordic nRF52 (Adafruit Feather nRF52), nRF51 (micro:bit)
+
+  Check forks for other architectures not listed here!
+
+- ### GNU Lesser General Public License
+
+  Adafruit_NeoPixel is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
+
+## Functions
+
+- begin()
+- updateLength()
+- updateType()
+- show()
+- delay_ns()
+- setPin()
+- setPixelColor()
+- fill()
+- ColorHSV()
+- getPixelColor()
+- setBrightness()
+- getBrightness()
+- clear()
+- gamma32()
+
+## Examples
+
+There are many examples implemented in this library. One of the examples is below. You can find other examples [here](https://github.com/adafruit/Adafruit_NeoPixel/tree/master/examples)
+
+### Simple
+
+```Cpp
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+  #include <avr/power.h>
+#endif
+#define PIN        6
+#define NUMPIXELS 16
+
+Adafruit_NeoPixel pixels(NUMPIXELS, PIN, NEO_GRB + NEO_KHZ800);
+#define DELAYVAL 500
+
+void setup() {
+#if defined(__AVR_ATtiny85__) && (F_CPU == 16000000)
+  clock_prescale_set(clock_div_1);
+#endif
+
+  pixels.begin();
+}
+
+void loop() {
+  pixels.clear();
+
+  for(int i=0; i<NUMPIXELS; i++) {
+
+    pixels.setPixelColor(i, pixels.Color(0, 150, 0));
+    pixels.show();
+    delay(DELAYVAL);
+  }
+}
+```
+
+## Contributing
+
+If you want to contribute to this project:
+
+- Report bugs and errors
+- Ask for enhancements
+- Create issues and pull requests
+- Tell others about this library
+- Contribute new protocols
+
+Please read [CONTRIBUTING.md](https://github.com/adafruit/Adafruit_NeoPixel/blob/master/CONTRIBUTING.md) for details on our code of conduct, and the process for submitting pull requests to us.
+
+### Roadmap
+
+The PRIME DIRECTIVE is to maintain backward compatibility with existing Arduino sketches -- many are hosted elsewhere and don't track changes here, some are in print and can never be changed!
+
+Please don't reformat code for the sake of reformatting code. The resulting large "visual diff" makes it impossible to untangle actual bug fixes from merely rearranged lines. (Exception for first item in wishlist below.)
+
+Things I'd Like To Do But There's No Official Timeline So Please Don't Count On Any Of This Ever Being Canonical:
+
+- For the show() function (with all the delicate pixel timing stuff), break out each architecture into separate source files rather than the current unmaintainable tangle of #ifdef statements!
+- Please don't use updateLength() or updateType() in new code. They should not have been implemented this way (use the C++ 'new' operator with the regular constructor instead) and are only sticking around because of the Prime Directive. setPin() is OK for now though, it's a trick we can use to 'recycle' pixel memory across multiple strips.
+- In the M0 and M4 code, use the hardware systick counter for bit timing rather than hand-tweaked NOPs (a temporary kludge at the time because I wasn't reading systick correctly). (As of 1.4.2, systick is used on M4 devices and it appears to be overclock-compatible. Not for M0 yet, which is why this item is still here.)
+- As currently written, brightness scaling is still a "destructive" operation -- pixel values are altered in RAM and the original value as set can't be accurately read back, only approximated, which has been confusing and frustrating to users. It was done this way at the time because NeoPixel timing is strict, AVR microcontrollers (all we had at the time) are limited, and assembly language is hard. All the 32-bit architectures should have no problem handling nondestructive brightness scaling -- calculating each byte immediately before it's sent out the wire, maintaining the original set value in RAM -- the work just hasn't been done. There's a fair chance even the AVR code could manage it with some intense focus. (The DotStar library achieves nondestructive brightness scaling because it doesn't have to manage data timing so carefully...every architecture, even ATtiny, just takes whatever cycles it needs for the multiply/shift operations.)
+
+## Credits
+
+This library is written by Phil "Paint Your Dragon" Burgess for Adafruit Industries, with contributions by PJRC, Michael Miller and other members of the open source community.
+
+## License
+
+Adafruit_NeoPixel is free software: you can redistribute it and/or  modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
+Adafruit_NeoPixel is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the [GNU Lesser General Public License](https://www.gnu.org/licenses/lgpl-3.0.en.html) for more details.
+You should have received a copy of the GNU Lesser General Public License along with NeoPixel.  If not, see [this](https://www.gnu.org/licenses/)

Adafruit_NeoPixel/esp.c

@@ -0,0 +1,178 @@
+// Implements the RMT peripheral on Espressif SoCs
+// Copyright (c) 2020 Lucian Copeland for Adafruit Industries
+
+/* Uses code from Espressif RGB LED Strip demo and drivers
+ * Copyright 2015-2020 Espressif Systems (Shanghai) PTE LTD
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#if defined(ESP32)
+
+#include <Arduino.h>
+#include "driver/rmt.h"
+
+#if defined(ESP_IDF_VERSION)
+#if ESP_IDF_VERSION >= ESP_IDF_VERSION_VAL(4, 0, 0)
+#define HAS_ESP_IDF_4
+#endif
+#endif
+
+// This code is adapted from the ESP-IDF v3.4 RMT "led_strip" example, altered
+// to work with the Arduino version of the ESP-IDF (3.2)
+
+#define WS2812_T0H_NS (400)
+#define WS2812_T0L_NS (850)
+#define WS2812_T1H_NS (800)
+#define WS2812_T1L_NS (450)
+
+#define WS2811_T0H_NS (500)
+#define WS2811_T0L_NS (2000)
+#define WS2811_T1H_NS (1200)
+#define WS2811_T1L_NS (1300)
+
+static uint32_t t0h_ticks = 0;
+static uint32_t t1h_ticks = 0;
+static uint32_t t0l_ticks = 0;
+static uint32_t t1l_ticks = 0;
+
+// Limit the number of RMT channels available for the Neopixels. Defaults to all
+// channels (8 on ESP32, 4 on ESP32-S2 and S3). Redefining this value will free
+// any channels with a higher number for other uses, such as IR send-and-recieve
+// libraries. Redefine as 1 to restrict Neopixels to only a single channel.
+#define ADAFRUIT_RMT_CHANNEL_MAX RMT_CHANNEL_MAX
+
+#define RMT_LL_HW_BASE  (&RMT)
+
+bool rmt_reserved_channels[ADAFRUIT_RMT_CHANNEL_MAX];
+
+static void IRAM_ATTR ws2812_rmt_adapter(const void *src, rmt_item32_t *dest, size_t src_size,
+        size_t wanted_num, size_t *translated_size, size_t *item_num)
+{
+    if (src == NULL || dest == NULL) {
+        *translated_size = 0;
+        *item_num = 0;
+        return;
+    }
+    const rmt_item32_t bit0 = {{{ t0h_ticks, 1, t0l_ticks, 0 }}}; //Logical 0
+    const rmt_item32_t bit1 = {{{ t1h_ticks, 1, t1l_ticks, 0 }}}; //Logical 1
+    size_t size = 0;
+    size_t num = 0;
+    uint8_t *psrc = (uint8_t *)src;
+    rmt_item32_t *pdest = dest;
+    while (size < src_size && num < wanted_num) {
+        for (int i = 0; i < 8; i++) {
+            // MSB first
+            if (*psrc & (1 << (7 - i))) {
+                pdest->val =  bit1.val;
+            } else {
+                pdest->val =  bit0.val;
+            }
+            num++;
+            pdest++;
+        }
+        size++;
+        psrc++;
+    }
+    *translated_size = size;
+    *item_num = num;
+}
+
+void espShow(uint8_t pin, uint8_t *pixels, uint32_t numBytes, boolean is800KHz) {
+    // Reserve channel
+    rmt_channel_t channel = ADAFRUIT_RMT_CHANNEL_MAX;
+    for (size_t i = 0; i < ADAFRUIT_RMT_CHANNEL_MAX; i++) {
+        if (!rmt_reserved_channels[i]) {
+            rmt_reserved_channels[i] = true;
+            channel = i;
+            break;
+        }
+    }
+    if (channel == ADAFRUIT_RMT_CHANNEL_MAX) {
+        // Ran out of channels!
+        return;
+    }
+
+#if defined(HAS_ESP_IDF_4)
+    rmt_config_t config = RMT_DEFAULT_CONFIG_TX(pin, channel);
+    config.clk_div = 2;
+#else
+    // Match default TX config from ESP-IDF version 3.4
+    rmt_config_t config = {
+        .rmt_mode = RMT_MODE_TX,
+        .channel = channel,
+        .gpio_num = pin,
+        .clk_div = 2,
+        .mem_block_num = 1,
+        .tx_config = {
+            .carrier_freq_hz = 38000,
+            .carrier_level = RMT_CARRIER_LEVEL_HIGH,
+            .idle_level = RMT_IDLE_LEVEL_LOW,
+            .carrier_duty_percent = 33,
+            .carrier_en = false,
+            .loop_en = false,
+            .idle_output_en = true,
+        }
+    };
+#endif
+    rmt_config(&config);
+    rmt_driver_install(config.channel, 0, 0);
+
+    // Convert NS timings to ticks
+    uint32_t counter_clk_hz = 0;
+
+#if defined(HAS_ESP_IDF_4)
+    rmt_get_counter_clock(channel, &counter_clk_hz);
+#else
+    // this emulates the rmt_get_counter_clock() function from ESP-IDF 3.4
+    if (RMT_LL_HW_BASE->conf_ch[config.channel].conf1.ref_always_on == RMT_BASECLK_REF) {
+        uint32_t div_cnt = RMT_LL_HW_BASE->conf_ch[config.channel].conf0.div_cnt;
+        uint32_t div = div_cnt == 0 ? 256 : div_cnt;
+        counter_clk_hz = REF_CLK_FREQ / (div);
+    } else {
+        uint32_t div_cnt = RMT_LL_HW_BASE->conf_ch[config.channel].conf0.div_cnt;
+        uint32_t div = div_cnt == 0 ? 256 : div_cnt;
+        counter_clk_hz = APB_CLK_FREQ / (div);
+    }
+#endif
+
+    // NS to tick converter
+    float ratio = (float)counter_clk_hz / 1e9;
+
+    if (is800KHz) {
+        t0h_ticks = (uint32_t)(ratio * WS2812_T0H_NS);
+        t0l_ticks = (uint32_t)(ratio * WS2812_T0L_NS);
+        t1h_ticks = (uint32_t)(ratio * WS2812_T1H_NS);
+        t1l_ticks = (uint32_t)(ratio * WS2812_T1L_NS);
+    } else {
+        t0h_ticks = (uint32_t)(ratio * WS2811_T0H_NS);
+        t0l_ticks = (uint32_t)(ratio * WS2811_T0L_NS);
+        t1h_ticks = (uint32_t)(ratio * WS2811_T1H_NS);
+        t1l_ticks = (uint32_t)(ratio * WS2811_T1L_NS);
+    }
+
+    // Initialize automatic timing translator
+    rmt_translator_init(config.channel, ws2812_rmt_adapter);
+
+    // Write and wait to finish
+    rmt_write_sample(config.channel, pixels, (size_t)numBytes, true);
+    rmt_wait_tx_done(config.channel, pdMS_TO_TICKS(100));
+
+    // Free channel again
+    rmt_driver_uninstall(config.channel);
+    rmt_reserved_channels[channel] = false;
+
+    gpio_set_direction(pin, GPIO_MODE_OUTPUT);
+}
+
+#endif

Adafruit_NeoPixel/esp8266.c

@@ -0,0 +1,86 @@
+// This is a mash-up of the Due show() code + insights from Michael Miller's
+// ESP8266 work for the NeoPixelBus library: github.com/Makuna/NeoPixelBus
+// Needs to be a separate .c file to enforce ICACHE_RAM_ATTR execution.
+
+#if defined(ESP8266)
+
+#include <Arduino.h>
+#ifdef ESP8266
+#include <eagle_soc.h>
+#endif
+
+static uint32_t _getCycleCount(void) __attribute__((always_inline));
+static inline uint32_t _getCycleCount(void) {
+  uint32_t ccount;
+  __asm__ __volatile__("rsr %0,ccount":"=a" (ccount));
+  return ccount;
+}
+
+#ifdef ESP8266
+IRAM_ATTR void espShow(
+ uint8_t pin, uint8_t *pixels, uint32_t numBytes, __attribute__((unused)) boolean is800KHz) {
+#else
+void espShow(
+ uint8_t pin, uint8_t *pixels, uint32_t numBytes, boolean is800KHz) {
+#endif
+
+#define CYCLES_800_T0H  (F_CPU / 2500001) // 0.4us
+#define CYCLES_800_T1H  (F_CPU / 1250001) // 0.8us
+#define CYCLES_800      (F_CPU /  800001) // 1.25us per bit
+#define CYCLES_400_T0H  (F_CPU / 2000000) // 0.5uS
+#define CYCLES_400_T1H  (F_CPU /  833333) // 1.2us
+#define CYCLES_400      (F_CPU /  400000) // 2.5us per bit
+
+  uint8_t *p, *end, pix, mask;
+  uint32_t t, time0, time1, period, c, startTime;
+
+#ifdef ESP8266
+  uint32_t pinMask;
+  pinMask   = _BV(pin);
+#endif
+
+  p         =  pixels;
+  end       =  p + numBytes;
+  pix       = *p++;
+  mask      = 0x80;
+  startTime = 0;
+
+#ifdef NEO_KHZ400
+  if(is800KHz) {
+#endif
+    time0  = CYCLES_800_T0H;
+    time1  = CYCLES_800_T1H;
+    period = CYCLES_800;
+#ifdef NEO_KHZ400
+  } else { // 400 KHz bitstream
+    time0  = CYCLES_400_T0H;
+    time1  = CYCLES_400_T1H;
+    period = CYCLES_400;
+  }
+#endif
+
+  for(t = time0;; t = time0) {
+    if(pix & mask) t = time1;                             // Bit high duration
+    while(((c = _getCycleCount()) - startTime) < period); // Wait for bit start
+#ifdef ESP8266
+    GPIO_REG_WRITE(GPIO_OUT_W1TS_ADDRESS, pinMask);       // Set high
+#else
+    gpio_set_level(pin, HIGH);
+#endif
+    startTime = c;                                        // Save start time
+    while(((c = _getCycleCount()) - startTime) < t);      // Wait high duration
+#ifdef ESP8266
+    GPIO_REG_WRITE(GPIO_OUT_W1TC_ADDRESS, pinMask);       // Set low
+#else
+    gpio_set_level(pin, LOW);
+#endif
+    if(!(mask >>= 1)) {                                   // Next bit/byte
+      if(p >= end) break;
+      pix  = *p++;
+      mask = 0x80;
+    }
+  }
+  while((_getCycleCount() - startTime) < period); // Wait for last bit
+}
+
+#endif // ESP8266

Adafruit_NeoPixel/examples/RGBWstrandtest/RGBWstrandtest.ino

@@ -0,0 +1,177 @@
+// NeoPixel test program showing use of the WHITE channel for RGBW
+// pixels only (won't look correct on regular RGB NeoPixel strips).
+
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+ #include <avr/power.h> // Required for 16 MHz Adafruit Trinket
+#endif
+
+// Which pin on the Arduino is connected to the NeoPixels?
+// On a Trinket or Gemma we suggest changing this to 1:
+#define LED_PIN     6
+
+// How many NeoPixels are attached to the Arduino?
+#define LED_COUNT  60
+
+// NeoPixel brightness, 0 (min) to 255 (max)
+#define BRIGHTNESS 50 // Set BRIGHTNESS to about 1/5 (max = 255)
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(LED_COUNT, LED_PIN, NEO_GRBW + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+void setup() {
+  // These lines are specifically to support the Adafruit Trinket 5V 16 MHz.
+  // Any other board, you can remove this part (but no harm leaving it):
+#if defined(__AVR_ATtiny85__) && (F_CPU == 16000000)
+  clock_prescale_set(clock_div_1);
+#endif
+  // END of Trinket-specific code.
+
+  strip.begin();           // INITIALIZE NeoPixel strip object (REQUIRED)
+  strip.show();            // Turn OFF all pixels ASAP
+  strip.setBrightness(BRIGHTNESS);
+}
+
+void loop() {
+  // Fill along the length of the strip in various colors...
+  colorWipe(strip.Color(255,   0,   0)     , 50); // Red
+  colorWipe(strip.Color(  0, 255,   0)     , 50); // Green
+  colorWipe(strip.Color(  0,   0, 255)     , 50); // Blue
+  colorWipe(strip.Color(  0,   0,   0, 255), 50); // True white (not RGB white)
+
+  whiteOverRainbow(75, 5);
+
+  pulseWhite(5);
+
+  rainbowFade2White(3, 3, 1);
+}
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait) {
+  for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+    strip.setPixelColor(i, color);         //  Set pixel's color (in RAM)
+    strip.show();                          //  Update strip to match
+    delay(wait);                           //  Pause for a moment
+  }
+}
+
+void whiteOverRainbow(int whiteSpeed, int whiteLength) {
+
+  if(whiteLength >= strip.numPixels()) whiteLength = strip.numPixels() - 1;
+
+  int      head          = whiteLength - 1;
+  int      tail          = 0;
+  int      loops         = 3;
+  int      loopNum       = 0;
+  uint32_t lastTime      = millis();
+  uint32_t firstPixelHue = 0;
+
+  for(;;) { // Repeat forever (or until a 'break' or 'return')
+    for(int i=0; i<strip.numPixels(); i++) {  // For each pixel in strip...
+      if(((i >= tail) && (i <= head)) ||      //  If between head & tail...
+         ((tail > head) && ((i >= tail) || (i <= head)))) {
+        strip.setPixelColor(i, strip.Color(0, 0, 0, 255)); // Set white
+      } else {                                             // else set rainbow
+        int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+        strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+      }
+    }
+
+    strip.show(); // Update strip with new contents
+    // There's no delay here, it just runs full-tilt until the timer and
+    // counter combination below runs out.
+
+    firstPixelHue += 40; // Advance just a little along the color wheel
+
+    if((millis() - lastTime) > whiteSpeed) { // Time to update head/tail?
+      if(++head >= strip.numPixels()) {      // Advance head, wrap around
+        head = 0;
+        if(++loopNum >= loops) return;
+      }
+      if(++tail >= strip.numPixels()) {      // Advance tail, wrap around
+        tail = 0;
+      }
+      lastTime = millis();                   // Save time of last movement
+    }
+  }
+}
+
+void pulseWhite(uint8_t wait) {
+  for(int j=0; j<256; j++) { // Ramp up from 0 to 255
+    // Fill entire strip with white at gamma-corrected brightness level 'j':
+    strip.fill(strip.Color(0, 0, 0, strip.gamma8(j)));
+    strip.show();
+    delay(wait);
+  }
+
+  for(int j=255; j>=0; j--) { // Ramp down from 255 to 0
+    strip.fill(strip.Color(0, 0, 0, strip.gamma8(j)));
+    strip.show();
+    delay(wait);
+  }
+}
+
+void rainbowFade2White(int wait, int rainbowLoops, int whiteLoops) {
+  int fadeVal=0, fadeMax=100;
+
+  // Hue of first pixel runs 'rainbowLoops' complete loops through the color
+  // wheel. Color wheel has a range of 65536 but it's OK if we roll over, so
+  // just count from 0 to rainbowLoops*65536, using steps of 256 so we
+  // advance around the wheel at a decent clip.
+  for(uint32_t firstPixelHue = 0; firstPixelHue < rainbowLoops*65536;
+    firstPixelHue += 256) {
+
+    for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+
+      // Offset pixel hue by an amount to make one full revolution of the
+      // color wheel (range of 65536) along the length of the strip
+      // (strip.numPixels() steps):
+      uint32_t pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+
+      // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+      // optionally add saturation and value (brightness) (each 0 to 255).
+      // Here we're using just the three-argument variant, though the
+      // second value (saturation) is a constant 255.
+      strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue, 255,
+        255 * fadeVal / fadeMax)));
+    }
+
+    strip.show();
+    delay(wait);
+
+    if(firstPixelHue < 65536) {                              // First loop,
+      if(fadeVal < fadeMax) fadeVal++;                       // fade in
+    } else if(firstPixelHue >= ((rainbowLoops-1) * 65536)) { // Last loop,
+      if(fadeVal > 0) fadeVal--;                             // fade out
+    } else {
+      fadeVal = fadeMax; // Interim loop, make sure fade is at max
+    }
+  }
+
+  for(int k=0; k<whiteLoops; k++) {
+    for(int j=0; j<256; j++) { // Ramp up 0 to 255
+      // Fill entire strip with white at gamma-corrected brightness level 'j':
+      strip.fill(strip.Color(0, 0, 0, strip.gamma8(j)));
+      strip.show();
+    }
+    delay(1000); // Pause 1 second
+    for(int j=255; j>=0; j--) { // Ramp down 255 to 0
+      strip.fill(strip.Color(0, 0, 0, strip.gamma8(j)));
+      strip.show();
+    }
+  }
+
+  delay(500); // Pause 1/2 second
+}

Adafruit_NeoPixel/examples/StrandtestArduinoBLE/StrandtestArduinoBLE.ino

@@ -0,0 +1,231 @@
+/****************************************************************************
+ * This example is based on StrandtestBLE example and adapts it to use 
+ * the new ArduinoBLE library.
+ * 
+ * https://github.com/arduino-libraries/ArduinoBLE
+ * 
+ * Supported boards:
+ *  Arduino MKR WiFi 1010, Arduino Uno WiFi Rev2 board, Arduino Nano 33 IoT,
+    Arduino Nano 33 BLE, or Arduino Nano 33 BLE Sense board.
+ * 
+ * You can use a generic BLE central app, like LightBlue (iOS and Android) or
+ * nRF Connect (Android), to interact with the services and characteristics
+ * created in this sketch.
+ * 
+ * This example code is in the public domain.
+ * 
+ */
+#include <Adafruit_NeoPixel.h>
+
+#define PIN 15 // Pin where NeoPixels are connected
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(64, PIN, NEO_GRB + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+// NEOPIXEL BEST PRACTICES for most reliable operation:
+// - Add 1000 uF CAPACITOR between NeoPixel strip's + and - connections.
+// - MINIMIZE WIRING LENGTH between microcontroller board and first pixel.
+// - NeoPixel strip's DATA-IN should pass through a 300-500 OHM RESISTOR.
+// - AVOID connecting NeoPixels on a LIVE CIRCUIT. If you must, ALWAYS
+//   connect GROUND (-) first, then +, then data.
+// - When using a 3.3V microcontroller with a 5V-powered NeoPixel strip,
+//   a LOGIC-LEVEL CONVERTER on the data line is STRONGLY RECOMMENDED.
+// (Skipping these may work OK on your workbench but can fail in the field)
+
+uint8_t rgb_values[3];
+
+#include <ArduinoBLE.h>
+
+BLEService ledService("19B10000-E8F2-537E-4F6C-D104768A1214"); // BLE LED Service
+
+// BLE LED Switch Characteristic - custom 128-bit UUID, read and writable by central
+BLEByteCharacteristic switchCharacteristic("19B10001-E8F2-537E-4F6C-D104768A1214", BLERead | BLEWrite);
+
+void setup()
+{
+    Serial.begin(115200);
+    Serial.println("Hello World!");
+
+    // custom services and characteristics can be added as well
+    // begin initialization
+    if (!BLE.begin())
+    {
+        Serial.println("starting BLE failed!");
+
+        while (1)
+            ;
+    }
+
+    Serial.print("Peripheral address: ");
+    Serial.println(BLE.address());
+
+    // set advertised local name and service UUID:
+    BLE.setLocalName("LED");
+    BLE.setAdvertisedService(ledService);
+
+    // add the characteristic to the service
+    ledService.addCharacteristic(switchCharacteristic);
+
+    // add service
+    BLE.addService(ledService);
+
+    // set the initial value for the characeristic:
+    switchCharacteristic.writeValue(0);
+
+    // start advertising
+    BLE.advertise();
+
+    strip.begin(); // INITIALIZE NeoPixel strip object (REQUIRED)
+    strip.show();  // Turn OFF all pixels ASAP
+
+    pinMode(PIN, OUTPUT);
+    digitalWrite(PIN, LOW);
+
+}
+
+void loop()
+{
+    BLEDevice central = BLE.central();
+
+    // if a central is connected to peripheral:
+    if (central)
+    {
+        Serial.print("Connected to central: ");
+        // print the central's MAC address:
+        Serial.println(central.address());
+
+        // while the central is still connected to peripheral:
+        while (central.connected())
+        {
+            // if the remote device wrote to the characteristic,
+            // use the value to control the LED:
+            if (switchCharacteristic.written())
+            {
+                switch (switchCharacteristic.value())
+                {
+                case 'a':
+                    colorWipe(strip.Color(255, 0, 0), 20); // Red
+                    break;
+                case 'b':
+                    colorWipe(strip.Color(0, 255, 0), 20); // Green
+                    break;
+                case 'c':
+                    colorWipe(strip.Color(0, 0, 255), 20); // Blue
+                    break;
+                case 'd':
+                    theaterChase(strip.Color(255, 0, 0), 20); // Red
+                    break;
+                case 'e':
+                    theaterChase(strip.Color(0, 255, 0), 20); // Green
+                    break;
+                case 'f':
+                    theaterChase(strip.Color(255, 0, 255), 20); // Cyan
+                    break;
+                case 'g':
+                    rainbow(10);
+                    break;
+                case 'h':
+                    theaterChaseRainbow(20);
+                    break;
+                }
+            }
+        }
+    }
+}
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait)
+{
+    for (int i = 0; i < strip.numPixels(); i++)
+    {                                  // For each pixel in strip...
+        strip.setPixelColor(i, color); //  Set pixel's color (in RAM)
+        strip.show();                  //  Update strip to match
+        delay(wait);                   //  Pause for a moment
+    }
+}
+
+// Theater-marquee-style chasing lights. Pass in a color (32-bit value,
+// a la strip.Color(r,g,b) as mentioned above), and a delay time (in ms)
+// between frames.
+void theaterChase(uint32_t color, int wait)
+{
+    for (int a = 0; a < 10; a++)
+    { // Repeat 10 times...
+        for (int b = 0; b < 3; b++)
+        {                  //  'b' counts from 0 to 2...
+            strip.clear(); //   Set all pixels in RAM to 0 (off)
+            // 'c' counts up from 'b' to end of strip in steps of 3...
+            for (int c = b; c < strip.numPixels(); c += 3)
+            {
+                strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+            }
+            strip.show(); // Update strip with new contents
+            delay(wait);  // Pause for a moment
+        }
+    }
+}
+
+// Rainbow cycle along whole strip. Pass delay time (in ms) between frames.
+void rainbow(int wait)
+{
+    // Hue of first pixel runs 5 complete loops through the color wheel.
+    // Color wheel has a range of 65536 but it's OK if we roll over, so
+    // just count from 0 to 5*65536. Adding 256 to firstPixelHue each time
+    // means we'll make 5*65536/256 = 1280 passes through this outer loop:
+    for (long firstPixelHue = 0; firstPixelHue < 5 * 65536; firstPixelHue += 256)
+    {
+        for (int i = 0; i < strip.numPixels(); i++)
+        { // For each pixel in strip...
+            // Offset pixel hue by an amount to make one full revolution of the
+            // color wheel (range of 65536) along the length of the strip
+            // (strip.numPixels() steps):
+            int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+            // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+            // optionally add saturation and value (brightness) (each 0 to 255).
+            // Here we're using just the single-argument hue variant. The result
+            // is passed through strip.gamma32() to provide 'truer' colors
+            // before assigning to each pixel:
+            strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+        }
+        strip.show(); // Update strip with new contents
+        delay(wait);  // Pause for a moment
+    }
+}
+
+// Rainbow-enhanced theater marquee. Pass delay time (in ms) between frames.
+void theaterChaseRainbow(int wait)
+{
+    int firstPixelHue = 0; // First pixel starts at red (hue 0)
+    for (int a = 0; a < 30; a++)
+    { // Repeat 30 times...
+        for (int b = 0; b < 3; b++)
+        {                  //  'b' counts from 0 to 2...
+            strip.clear(); //   Set all pixels in RAM to 0 (off)
+            // 'c' counts up from 'b' to end of strip in increments of 3...
+            for (int c = b; c < strip.numPixels(); c += 3)
+            {
+                // hue of pixel 'c' is offset by an amount to make one full
+                // revolution of the color wheel (range 65536) along the length
+                // of the strip (strip.numPixels() steps):
+                int hue = firstPixelHue + c * 65536L / strip.numPixels();
+                uint32_t color = strip.gamma32(strip.ColorHSV(hue)); // hue -> RGB
+                strip.setPixelColor(c, color);                       // Set pixel 'c' to value 'color'
+            }
+            strip.show();                // Update strip with new contents
+            delay(wait);                 // Pause for a moment
+            firstPixelHue += 65536 / 90; // One cycle of color wheel over 90 frames
+        }
+    }
+}

Adafruit_NeoPixel/examples/StrandtestArduinoBLECallback/StrandtestArduinoBLECallback.ino

@@ -0,0 +1,239 @@
+/****************************************************************************
+ * This example is based on StrandtestArduinoBLE example to make use of
+ * callbacks features of the ArduinoBLE library.
+ * 
+ * https://github.com/arduino-libraries/ArduinoBLE 
+ * 
+ * Supported boards:
+ *  Arduino MKR WiFi 1010, Arduino Uno WiFi Rev2 board, Arduino Nano 33 IoT,
+    Arduino Nano 33 BLE, or Arduino Nano 33 BLE Sense board.
+ * 
+ * You can use a generic BLE central app, like LightBlue (iOS and Android) or
+ * nRF Connect (Android), to interact with the services and characteristics
+ * created in this sketch.
+ * 
+ * This example code is in the public domain.
+ * 
+ */
+#include <Adafruit_NeoPixel.h>
+
+#define PIN 15 // Pin where NeoPixels are connected
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(64, PIN, NEO_GRB + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+// NEOPIXEL BEST PRACTICES for most reliable operation:
+// - Add 1000 uF CAPACITOR between NeoPixel strip's + and - connections.
+// - MINIMIZE WIRING LENGTH between microcontroller board and first pixel.
+// - NeoPixel strip's DATA-IN should pass through a 300-500 OHM RESISTOR.
+// - AVOID connecting NeoPixels on a LIVE CIRCUIT. If you must, ALWAYS
+//   connect GROUND (-) first, then +, then data.
+// - When using a 3.3V microcontroller with a 5V-powered NeoPixel strip,
+//   a LOGIC-LEVEL CONVERTER on the data line is STRONGLY RECOMMENDED.
+// (Skipping these may work OK on your workbench but can fail in the field)
+
+uint8_t rgb_values[3];
+
+#include <ArduinoBLE.h>
+
+BLEService ledService("19B10000-E8F2-537E-4F6C-D104768A1214"); // BLE LED Service
+
+// BLE LED Switch Characteristic - custom 128-bit UUID, read and writable by central
+BLEByteCharacteristic switchCharacteristic("19B10001-E8F2-537E-4F6C-D104768A1214", BLERead | BLEWrite);
+
+void setup()
+{
+    Serial.begin(115200);
+    Serial.println("Hello World!");
+
+    // custom services and characteristics can be added as well
+    // begin initialization
+    if (!BLE.begin())
+    {
+        Serial.println("starting BLE failed!");
+
+        while (1)
+            ;
+    }
+
+    Serial.print("Peripheral address: ");
+    Serial.println(BLE.address());
+
+    // set advertised local name and service UUID:
+    BLE.setLocalName("LEDCallback");
+    BLE.setAdvertisedService(ledService);
+
+    // add the characteristic to the service
+    ledService.addCharacteristic(switchCharacteristic);
+
+    // add service
+    BLE.addService(ledService);
+    // assign event handlers for connected, disconnected to peripheral
+    BLE.setEventHandler(BLEConnected, blePeripheralConnectHandler);
+    BLE.setEventHandler(BLEDisconnected, blePeripheralDisconnectHandler);
+
+    // assign event handlers for characteristic
+    switchCharacteristic.setEventHandler(BLEWritten, switchCharacteristicWritten);
+    // set the initial value for the characeristic:
+    switchCharacteristic.writeValue(0);
+
+    // start advertising
+    BLE.advertise();
+
+    strip.begin(); // INITIALIZE NeoPixel strip object (REQUIRED)
+    strip.show();  // Turn OFF all pixels ASAP
+
+    pinMode(PIN, OUTPUT);
+    digitalWrite(PIN, LOW);
+}
+
+void loop()
+{
+    // poll for BLE events
+    BLE.poll();
+}
+
+void blePeripheralConnectHandler(BLEDevice central)
+{
+    // central connected event handler
+    Serial.print("Connected event, central: ");
+    Serial.println(central.address());
+}
+
+void blePeripheralDisconnectHandler(BLEDevice central)
+{
+    // central disconnected event handler
+    Serial.print("Disconnected event, central: ");
+    Serial.println(central.address());
+}
+
+void switchCharacteristicWritten(BLEDevice central, BLECharacteristic characteristic)
+{
+    // central wrote new value to characteristic, update LED
+    Serial.print("Characteristic event, written: ");
+
+    switch (switchCharacteristic.value())
+    {
+    case 'a':
+        colorWipe(strip.Color(255, 0, 0), 20); // Red
+        break;
+    case 'b':
+        colorWipe(strip.Color(0, 255, 0), 20); // Green
+        break;
+    case 'c':
+        colorWipe(strip.Color(0, 0, 255), 20); // Blue
+        break;
+    case 'd':
+        theaterChase(strip.Color(255, 0, 0), 20); // Red
+        break;
+    case 'e':
+        theaterChase(strip.Color(0, 255, 0), 20); // Green
+        break;
+    case 'f':
+        theaterChase(strip.Color(255, 0, 255), 20); // Cyan
+        break;
+    case 'g':
+        rainbow(10);
+        break;
+    case 'h':
+        theaterChaseRainbow(20);
+        break;
+    }
+}
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait)
+{
+    for (int i = 0; i < strip.numPixels(); i++)
+    {                                  // For each pixel in strip...
+        strip.setPixelColor(i, color); //  Set pixel's color (in RAM)
+        strip.show();                  //  Update strip to match
+        delay(wait);                   //  Pause for a moment
+    }
+}
+
+// Theater-marquee-style chasing lights. Pass in a color (32-bit value,
+// a la strip.Color(r,g,b) as mentioned above), and a delay time (in ms)
+// between frames.
+void theaterChase(uint32_t color, int wait)
+{
+    for (int a = 0; a < 10; a++)
+    { // Repeat 10 times...
+        for (int b = 0; b < 3; b++)
+        {                  //  'b' counts from 0 to 2...
+            strip.clear(); //   Set all pixels in RAM to 0 (off)
+            // 'c' counts up from 'b' to end of strip in steps of 3...
+            for (int c = b; c < strip.numPixels(); c += 3)
+            {
+                strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+            }
+            strip.show(); // Update strip with new contents
+            delay(wait);  // Pause for a moment
+        }
+    }
+}
+
+// Rainbow cycle along whole strip. Pass delay time (in ms) between frames.
+void rainbow(int wait)
+{
+    // Hue of first pixel runs 5 complete loops through the color wheel.
+    // Color wheel has a range of 65536 but it's OK if we roll over, so
+    // just count from 0 to 5*65536. Adding 256 to firstPixelHue each time
+    // means we'll make 5*65536/256 = 1280 passes through this outer loop:
+    for (long firstPixelHue = 0; firstPixelHue < 5 * 65536; firstPixelHue += 256)
+    {
+        for (int i = 0; i < strip.numPixels(); i++)
+        { // For each pixel in strip...
+            // Offset pixel hue by an amount to make one full revolution of the
+            // color wheel (range of 65536) along the length of the strip
+            // (strip.numPixels() steps):
+            int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+            // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+            // optionally add saturation and value (brightness) (each 0 to 255).
+            // Here we're using just the single-argument hue variant. The result
+            // is passed through strip.gamma32() to provide 'truer' colors
+            // before assigning to each pixel:
+            strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+        }
+        strip.show(); // Update strip with new contents
+        delay(wait);  // Pause for a moment
+    }
+}
+
+// Rainbow-enhanced theater marquee. Pass delay time (in ms) between frames.
+void theaterChaseRainbow(int wait)
+{
+    int firstPixelHue = 0; // First pixel starts at red (hue 0)
+    for (int a = 0; a < 30; a++)
+    { // Repeat 30 times...
+        for (int b = 0; b < 3; b++)
+        {                  //  'b' counts from 0 to 2...
+            strip.clear(); //   Set all pixels in RAM to 0 (off)
+            // 'c' counts up from 'b' to end of strip in increments of 3...
+            for (int c = b; c < strip.numPixels(); c += 3)
+            {
+                // hue of pixel 'c' is offset by an amount to make one full
+                // revolution of the color wheel (range 65536) along the length
+                // of the strip (strip.numPixels() steps):
+                int hue = firstPixelHue + c * 65536L / strip.numPixels();
+                uint32_t color = strip.gamma32(strip.ColorHSV(hue)); // hue -> RGB
+                strip.setPixelColor(c, color);                       // Set pixel 'c' to value 'color'
+            }
+            strip.show();                // Update strip with new contents
+            delay(wait);                 // Pause for a moment
+            firstPixelHue += 65536 / 90; // One cycle of color wheel over 90 frames
+        }
+    }
+}

Adafruit_NeoPixel/examples/StrandtestBLE/BLESerial.cpp

@@ -0,0 +1,133 @@
+#include "BLESerial.h"
+
+// #define BLE_SERIAL_DEBUG
+
+BLESerial* BLESerial::_instance = NULL;
+
+BLESerial::BLESerial(unsigned char req, unsigned char rdy, unsigned char rst) :
+  BLEPeripheral(req, rdy, rst)
+{
+  this->_txCount = 0;
+  this->_rxHead = this->_rxTail = 0;
+  this->_flushed = 0;
+  BLESerial::_instance = this;
+
+  addAttribute(this->_uartService);
+  addAttribute(this->_uartNameDescriptor);
+  setAdvertisedServiceUuid(this->_uartService.uuid());
+  addAttribute(this->_rxCharacteristic);
+  addAttribute(this->_rxNameDescriptor);
+  this->_rxCharacteristic.setEventHandler(BLEWritten, BLESerial::_received);
+  addAttribute(this->_txCharacteristic);
+  addAttribute(this->_txNameDescriptor);
+}
+
+void BLESerial::begin(...) {
+  BLEPeripheral::begin();
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.println(F("BLESerial::begin()"));
+  #endif
+}
+
+void BLESerial::poll() {
+  if (millis() < this->_flushed + 100) {
+    BLEPeripheral::poll();
+  } else {
+    flush();
+  }
+}
+
+void BLESerial::end() {
+  this->_rxCharacteristic.setEventHandler(BLEWritten, NULL);
+  this->_rxHead = this->_rxTail = 0;
+  flush();
+  BLEPeripheral::disconnect();
+}
+
+int BLESerial::available(void) {
+  BLEPeripheral::poll();
+  int retval = (this->_rxHead - this->_rxTail + sizeof(this->_rxBuffer)) % sizeof(this->_rxBuffer);
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::available() = "));
+    Serial.println(retval);
+  #endif
+  return retval;
+}
+
+int BLESerial::peek(void) {
+  BLEPeripheral::poll();
+  if (this->_rxTail == this->_rxHead) return -1;
+  uint8_t byte = this->_rxBuffer[this->_rxTail];
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::peek() = "));
+    Serial.print((char) byte);
+    Serial.print(F(" 0x"));
+    Serial.println(byte, HEX);
+  #endif
+  return byte;
+}
+
+int BLESerial::read(void) {
+  BLEPeripheral::poll();
+  if (this->_rxTail == this->_rxHead) return -1;
+  this->_rxTail = (this->_rxTail + 1) % sizeof(this->_rxBuffer);
+  uint8_t byte = this->_rxBuffer[this->_rxTail];
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::read() = "));
+    Serial.print((char) byte);
+    Serial.print(F(" 0x"));
+    Serial.println(byte, HEX);
+  #endif
+  return byte;
+}
+
+void BLESerial::flush(void) {
+  if (this->_txCount == 0) return;
+  this->_txCharacteristic.setValue(this->_txBuffer, this->_txCount);
+  this->_flushed = millis();
+  this->_txCount = 0;
+  BLEPeripheral::poll();
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.println(F("BLESerial::flush()"));
+  #endif
+}
+
+size_t BLESerial::write(uint8_t byte) {
+  BLEPeripheral::poll();
+  if (this->_txCharacteristic.subscribed() == false) return 0;
+  this->_txBuffer[this->_txCount++] = byte;
+  if (this->_txCount == sizeof(this->_txBuffer)) flush();
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::write("));
+    Serial.print((char) byte);
+    Serial.print(F(" 0x"));
+    Serial.print(byte, HEX);
+    Serial.println(F(") = 1"));
+  #endif
+  return 1;
+}
+
+BLESerial::operator bool() {
+  bool retval = BLEPeripheral::connected();
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::operator bool() = "));
+    Serial.println(retval);
+  #endif
+  return retval;
+}
+
+void BLESerial::_received(const uint8_t* data, size_t size) {
+  for (int i = 0; i < size; i++) {
+    this->_rxHead = (this->_rxHead + 1) % sizeof(this->_rxBuffer);
+    this->_rxBuffer[this->_rxHead] = data[i];
+  }
+  #ifdef BLE_SERIAL_DEBUG
+    Serial.print(F("BLESerial::received("));
+    for (int i = 0; i < size; i++) Serial.print((char) data[i]);
+    Serial.println(F(")"));
+  #endif
+}
+
+void BLESerial::_received(BLECentral& /*central*/, BLECharacteristic& rxCharacteristic) {
+  BLESerial::_instance->_received(rxCharacteristic.value(), rxCharacteristic.valueLength());
+}

Adafruit_NeoPixel/examples/StrandtestBLE/BLESerial.h

@@ -0,0 +1,46 @@
+#ifndef _BLE_SERIAL_H_
+#define _BLE_SERIAL_H_
+
+#include <Arduino.h>
+#include <BLEPeripheral.h>
+
+class BLESerial : public BLEPeripheral, public Stream
+{
+  public:
+    BLESerial(unsigned char req, unsigned char rdy, unsigned char rst);
+
+    void begin(...);
+    void poll();
+    void end();
+
+    virtual int available(void);
+    virtual int peek(void);
+    virtual int read(void);
+    virtual void flush(void);
+    virtual size_t write(uint8_t byte);
+    using Print::write;
+    virtual operator bool();
+
+  private:
+    unsigned long _flushed;
+    static BLESerial* _instance;
+
+    size_t _rxHead;
+    size_t _rxTail;
+    size_t _rxCount() const;
+    uint8_t _rxBuffer[BLE_ATTRIBUTE_MAX_VALUE_LENGTH];
+    size_t _txCount;
+    uint8_t _txBuffer[BLE_ATTRIBUTE_MAX_VALUE_LENGTH];
+
+    BLEService _uartService = BLEService("6E400001-B5A3-F393-E0A9-E50E24DCCA9E");
+    BLEDescriptor _uartNameDescriptor = BLEDescriptor("2901", "UART");
+    BLECharacteristic _rxCharacteristic = BLECharacteristic("6E400002-B5A3-F393-E0A9-E50E24DCCA9E", BLEWriteWithoutResponse, BLE_ATTRIBUTE_MAX_VALUE_LENGTH);
+    BLEDescriptor _rxNameDescriptor = BLEDescriptor("2901", "RX - Receive Data (Write)");
+    BLECharacteristic _txCharacteristic = BLECharacteristic("6E400003-B5A3-F393-E0A9-E50E24DCCA9E", BLENotify, BLE_ATTRIBUTE_MAX_VALUE_LENGTH);
+    BLEDescriptor _txNameDescriptor = BLEDescriptor("2901", "TX - Transfer Data (Notify)");
+
+    void _received(const uint8_t* data, size_t size);
+    static void _received(BLECentral& /*central*/, BLECharacteristic& rxCharacteristic);
+};
+
+#endif

Adafruit_NeoPixel/examples/StrandtestBLE/StrandtestBLE.ino

@@ -0,0 +1,192 @@
+/****************************************************************************
+ * This example was developed by the Hackerspace San Salvador to demonstrate
+ * the simultaneous use of the NeoPixel library and the Bluetooth SoftDevice.
+ * To compile this example you'll need to add support for the NRF52 based
+ * following the instructions at:
+ *  https://github.com/sandeepmistry/arduino-nRF5
+ * Or adding the following URL to the board manager URLs on Arduino preferences:
+ *  https://sandeepmistry.github.io/arduino-nRF5/package_nRF5_boards_index.json
+ * Then you can install the BLEPeripheral library avaiable at:
+ *  https://github.com/sandeepmistry/arduino-BLEPeripheral
+ * To test it, compile this example and use the UART module from the nRF
+ * Toolbox App for Android. Edit the interface and send the characters
+ * 'a' to 'i' to switch the animation.
+ * There is a delay because this example blocks the thread of execution but
+ * the change will be shown after the current animation ends. (This might
+ * take a couple of seconds)
+ * For more info write us at: info _at- teubi.co
+ */
+#include <SPI.h>
+#include <BLEPeripheral.h>
+#include "BLESerial.h"
+#include <Adafruit_NeoPixel.h>
+
+#define PIN 15 // Pin where NeoPixels are connected
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(64, PIN, NEO_GRB + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+// NEOPIXEL BEST PRACTICES for most reliable operation:
+// - Add 1000 uF CAPACITOR between NeoPixel strip's + and - connections.
+// - MINIMIZE WIRING LENGTH between microcontroller board and first pixel.
+// - NeoPixel strip's DATA-IN should pass through a 300-500 OHM RESISTOR.
+// - AVOID connecting NeoPixels on a LIVE CIRCUIT. If you must, ALWAYS
+//   connect GROUND (-) first, then +, then data.
+// - When using a 3.3V microcontroller with a 5V-powered NeoPixel strip,
+//   a LOGIC-LEVEL CONVERTER on the data line is STRONGLY RECOMMENDED.
+// (Skipping these may work OK on your workbench but can fail in the field)
+
+// define pins (varies per shield/board)
+#define BLE_REQ   10
+#define BLE_RDY   2
+#define BLE_RST   9
+
+// create ble serial instance, see pinouts above
+BLESerial BLESerial(BLE_REQ, BLE_RDY, BLE_RST);
+
+uint8_t current_state = 0;
+uint8_t rgb_values[3];
+
+void setup() {
+  Serial.begin(115200);
+  Serial.println("Hello World!");
+  // custom services and characteristics can be added as well
+  BLESerial.setLocalName("UART_HS");
+  BLESerial.begin();
+
+  strip.begin(); // INITIALIZE NeoPixel strip object (REQUIRED)
+  strip.show();  // Turn OFF all pixels ASAP
+
+  //pinMode(PIN, OUTPUT);
+  //digitalWrite(PIN, LOW);
+
+  current_state = 'a';
+}
+
+void loop() {
+  while(BLESerial.available()) {
+    uint8_t character = BLESerial.read();
+    switch(character) {
+      case 'a':
+      case 'b':
+      case 'c':
+      case 'd':
+      case 'e':
+      case 'f':
+      case 'g':
+      case 'h':
+        current_state = character;
+        break;
+    };
+  }
+  switch(current_state) {
+    case 'a':
+      colorWipe(strip.Color(255,   0,   0), 20);    // Red
+      break;
+    case 'b':
+      colorWipe(strip.Color(  0, 255,   0), 20);    // Green
+      break;
+    case 'c':
+      colorWipe(strip.Color(  0,   0, 255), 20);    // Blue
+      break;
+    case 'd':
+      theaterChase(strip.Color(255,   0,   0), 20); // Red
+      break;
+    case 'e':
+      theaterChase(strip.Color(  0, 255,   0), 20); // Green
+      break;
+    case 'f':
+      theaterChase(strip.Color(255,   0, 255), 20); // Cyan
+      break;
+    case 'g':
+      rainbow(10);
+      break;
+    case 'h':
+      theaterChaseRainbow(20);
+      break;
+  }
+}
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait) {
+  for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+    strip.setPixelColor(i, color);         //  Set pixel's color (in RAM)
+    strip.show();                          //  Update strip to match
+    delay(wait);                           //  Pause for a moment
+  }
+}
+
+// Theater-marquee-style chasing lights. Pass in a color (32-bit value,
+// a la strip.Color(r,g,b) as mentioned above), and a delay time (in ms)
+// between frames.
+void theaterChase(uint32_t color, int wait) {
+  for(int a=0; a<10; a++) {  // Repeat 10 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in steps of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show(); // Update strip with new contents
+      delay(wait);  // Pause for a moment
+    }
+  }
+}
+
+// Rainbow cycle along whole strip. Pass delay time (in ms) between frames.
+void rainbow(int wait) {
+  // Hue of first pixel runs 5 complete loops through the color wheel.
+  // Color wheel has a range of 65536 but it's OK if we roll over, so
+  // just count from 0 to 5*65536. Adding 256 to firstPixelHue each time
+  // means we'll make 5*65536/256 = 1280 passes through this outer loop:
+  for(long firstPixelHue = 0; firstPixelHue < 5*65536; firstPixelHue += 256) {
+    for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+      // Offset pixel hue by an amount to make one full revolution of the
+      // color wheel (range of 65536) along the length of the strip
+      // (strip.numPixels() steps):
+      int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+      // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+      // optionally add saturation and value (brightness) (each 0 to 255).
+      // Here we're using just the single-argument hue variant. The result
+      // is passed through strip.gamma32() to provide 'truer' colors
+      // before assigning to each pixel:
+      strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+    }
+    strip.show(); // Update strip with new contents
+    delay(wait);  // Pause for a moment
+  }
+}
+
+// Rainbow-enhanced theater marquee. Pass delay time (in ms) between frames.
+void theaterChaseRainbow(int wait) {
+  int firstPixelHue = 0;     // First pixel starts at red (hue 0)
+  for(int a=0; a<30; a++) {  // Repeat 30 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in increments of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        // hue of pixel 'c' is offset by an amount to make one full
+        // revolution of the color wheel (range 65536) along the length
+        // of the strip (strip.numPixels() steps):
+        int      hue   = firstPixelHue + c * 65536L / strip.numPixels();
+        uint32_t color = strip.gamma32(strip.ColorHSV(hue)); // hue -> RGB
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show();                // Update strip with new contents
+      delay(wait);                 // Pause for a moment
+      firstPixelHue += 65536 / 90; // One cycle of color wheel over 90 frames
+    }
+  }
+}

Adafruit_NeoPixel/examples/buttoncycler/buttoncycler.ino

@@ -0,0 +1,164 @@
+// Simple demonstration on using an input device to trigger changes on your
+// NeoPixels. Wire a momentary push button to connect from ground to a
+// digital IO pin. When the button is pressed it will change to a new pixel
+// animation. Initial state has all pixels off -- press the button once to
+// start the first animation. As written, the button does not interrupt an
+// animation in-progress, it works only when idle.
+
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+ #include <avr/power.h> // Required for 16 MHz Adafruit Trinket
+#endif
+
+// Digital IO pin connected to the button. This will be driven with a
+// pull-up resistor so the switch pulls the pin to ground momentarily.
+// On a high -> low transition the button press logic will execute.
+#define BUTTON_PIN   2
+
+#define PIXEL_PIN    6  // Digital IO pin connected to the NeoPixels.
+
+#define PIXEL_COUNT 16  // Number of NeoPixels
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(PIXEL_COUNT, PIXEL_PIN, NEO_GRB + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+boolean oldState = HIGH;
+int     mode     = 0;    // Currently-active animation mode, 0-9
+
+void setup() {
+  pinMode(BUTTON_PIN, INPUT_PULLUP);
+  strip.begin(); // Initialize NeoPixel strip object (REQUIRED)
+  strip.show();  // Initialize all pixels to 'off'
+}
+
+void loop() {
+  // Get current button state.
+  boolean newState = digitalRead(BUTTON_PIN);
+
+  // Check if state changed from high to low (button press).
+  if((newState == LOW) && (oldState == HIGH)) {
+    // Short delay to debounce button.
+    delay(20);
+    // Check if button is still low after debounce.
+    newState = digitalRead(BUTTON_PIN);
+    if(newState == LOW) {      // Yes, still low
+      if(++mode > 8) mode = 0; // Advance to next mode, wrap around after #8
+      switch(mode) {           // Start the new animation...
+        case 0:
+          colorWipe(strip.Color(  0,   0,   0), 50);    // Black/off
+          break;
+        case 1:
+          colorWipe(strip.Color(255,   0,   0), 50);    // Red
+          break;
+        case 2:
+          colorWipe(strip.Color(  0, 255,   0), 50);    // Green
+          break;
+        case 3:
+          colorWipe(strip.Color(  0,   0, 255), 50);    // Blue
+          break;
+        case 4:
+          theaterChase(strip.Color(127, 127, 127), 50); // White
+          break;
+        case 5:
+          theaterChase(strip.Color(127,   0,   0), 50); // Red
+          break;
+        case 6:
+          theaterChase(strip.Color(  0,   0, 127), 50); // Blue
+          break;
+        case 7:
+          rainbow(10);
+          break;
+        case 8:
+          theaterChaseRainbow(50);
+          break;
+      }
+    }
+  }
+
+  // Set the last-read button state to the old state.
+  oldState = newState;
+}
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait) {
+  for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+    strip.setPixelColor(i, color);         //  Set pixel's color (in RAM)
+    strip.show();                          //  Update strip to match
+    delay(wait);                           //  Pause for a moment
+  }
+}
+
+// Theater-marquee-style chasing lights. Pass in a color (32-bit value,
+// a la strip.Color(r,g,b) as mentioned above), and a delay time (in ms)
+// between frames.
+void theaterChase(uint32_t color, int wait) {
+  for(int a=0; a<10; a++) {  // Repeat 10 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in steps of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show(); // Update strip with new contents
+      delay(wait);  // Pause for a moment
+    }
+  }
+}
+
+// Rainbow cycle along whole strip. Pass delay time (in ms) between frames.
+void rainbow(int wait) {
+  // Hue of first pixel runs 3 complete loops through the color wheel.
+  // Color wheel has a range of 65536 but it's OK if we roll over, so
+  // just count from 0 to 3*65536. Adding 256 to firstPixelHue each time
+  // means we'll make 3*65536/256 = 768 passes through this outer loop:
+  for(long firstPixelHue = 0; firstPixelHue < 3*65536; firstPixelHue += 256) {
+    for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+      // Offset pixel hue by an amount to make one full revolution of the
+      // color wheel (range of 65536) along the length of the strip
+      // (strip.numPixels() steps):
+      int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+      // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+      // optionally add saturation and value (brightness) (each 0 to 255).
+      // Here we're using just the single-argument hue variant. The result
+      // is passed through strip.gamma32() to provide 'truer' colors
+      // before assigning to each pixel:
+      strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+    }
+    strip.show(); // Update strip with new contents
+    delay(wait);  // Pause for a moment
+  }
+}
+
+// Rainbow-enhanced theater marquee. Pass delay time (in ms) between frames.
+void theaterChaseRainbow(int wait) {
+  int firstPixelHue = 0;     // First pixel starts at red (hue 0)
+  for(int a=0; a<30; a++) {  // Repeat 30 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in increments of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        // hue of pixel 'c' is offset by an amount to make one full
+        // revolution of the color wheel (range 65536) along the length
+        // of the strip (strip.numPixels() steps):
+        int      hue   = firstPixelHue + c * 65536L / strip.numPixels();
+        uint32_t color = strip.gamma32(strip.ColorHSV(hue)); // hue -> RGB
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show();                // Update strip with new contents
+      delay(wait);                 // Pause for a moment
+      firstPixelHue += 65536 / 90; // One cycle of color wheel over 90 frames
+    }
+  }
+}

Adafruit_NeoPixel/examples/simple/simple.ino

@@ -0,0 +1,50 @@
+// NeoPixel Ring simple sketch (c) 2013 Shae Erisson
+// Released under the GPLv3 license to match the rest of the
+// Adafruit NeoPixel library
+
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+ #include <avr/power.h> // Required for 16 MHz Adafruit Trinket
+#endif
+
+// Which pin on the Arduino is connected to the NeoPixels?
+#define PIN        6 // On Trinket or Gemma, suggest changing this to 1
+
+// How many NeoPixels are attached to the Arduino?
+#define NUMPIXELS 16 // Popular NeoPixel ring size
+
+// When setting up the NeoPixel library, we tell it how many pixels,
+// and which pin to use to send signals. Note that for older NeoPixel
+// strips you might need to change the third parameter -- see the
+// strandtest example for more information on possible values.
+Adafruit_NeoPixel pixels(NUMPIXELS, PIN, NEO_GRB + NEO_KHZ800);
+
+#define DELAYVAL 500 // Time (in milliseconds) to pause between pixels
+
+void setup() {
+  // These lines are specifically to support the Adafruit Trinket 5V 16 MHz.
+  // Any other board, you can remove this part (but no harm leaving it):
+#if defined(__AVR_ATtiny85__) && (F_CPU == 16000000)
+  clock_prescale_set(clock_div_1);
+#endif
+  // END of Trinket-specific code.
+
+  pixels.begin(); // INITIALIZE NeoPixel strip object (REQUIRED)
+}
+
+void loop() {
+  pixels.clear(); // Set all pixel colors to 'off'
+
+  // The first NeoPixel in a strand is #0, second is 1, all the way up
+  // to the count of pixels minus one.
+  for(int i=0; i<NUMPIXELS; i++) { // For each pixel...
+
+    // pixels.Color() takes RGB values, from 0,0,0 up to 255,255,255
+    // Here we're using a moderately bright green color:
+    pixels.setPixelColor(i, pixels.Color(0, 150, 0));
+
+    pixels.show();   // Send the updated pixel colors to the hardware.
+
+    delay(DELAYVAL); // Pause before next pass through loop
+  }
+}

Adafruit_NeoPixel/examples/simple_new_operator/simple_new_operator.ino

@@ -0,0 +1,67 @@
+// NeoPixel Ring simple sketch (c) 2013 Shae Erisson
+// Released under the GPLv3 license to match the rest of the
+// Adafruit NeoPixel library
+// This sketch shows use of the "new" operator with Adafruit_NeoPixel.
+// It's helpful if you don't know NeoPixel settings at compile time or
+// just want to store this settings in EEPROM or a file on an SD card.
+
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+ #include <avr/power.h> // Required for 16 MHz Adafruit Trinket
+#endif
+
+// Which pin on the Arduino is connected to the NeoPixels?
+int pin         =  6; // On Trinket or Gemma, suggest changing this to 1
+
+// How many NeoPixels are attached to the Arduino?
+int numPixels   = 16; // Popular NeoPixel ring size
+
+// NeoPixel color format & data rate. See the strandtest example for
+// information on possible values.
+int pixelFormat = NEO_GRB + NEO_KHZ800;
+
+// Rather than declaring the whole NeoPixel object here, we just create
+// a pointer for one, which we'll then allocate later...
+Adafruit_NeoPixel *pixels;
+
+#define DELAYVAL 500 // Time (in milliseconds) to pause between pixels
+
+void setup() {
+  // These lines are specifically to support the Adafruit Trinket 5V 16 MHz.
+  // Any other board, you can remove this part (but no harm leaving it):
+#if defined(__AVR_ATtiny85__) && (F_CPU == 16000000)
+  clock_prescale_set(clock_div_1);
+#endif
+  // END of Trinket-specific code.
+  
+  // Right about here is where we could read 'pin', 'numPixels' and/or
+  // 'pixelFormat' from EEPROM or a file on SD or whatever. This is a simple
+  // example and doesn't do that -- those variables are just set to fixed
+  // values at the top of this code -- but this is where it would happen.
+
+  // Then create a new NeoPixel object dynamically with these values:
+  pixels = new Adafruit_NeoPixel(numPixels, pin, pixelFormat);
+
+  // Going forward from here, code works almost identically to any other
+  // NeoPixel example, but instead of the dot operator on function calls
+  // (e.g. pixels.begin()), we instead use pointer indirection (->) like so:
+  pixels->begin(); // INITIALIZE NeoPixel strip object (REQUIRED)
+  // You'll see more of this in the loop() function below.
+}
+
+void loop() {
+  pixels->clear(); // Set all pixel colors to 'off'
+
+  // The first NeoPixel in a strand is #0, second is 1, all the way up
+  // to the count of pixels minus one.
+  for(int i=0; i<numPixels; i++) { // For each pixel...
+
+    // pixels->Color() takes RGB values, from 0,0,0 up to 255,255,255
+    // Here we're using a moderately bright green color:
+    pixels->setPixelColor(i, pixels->Color(0, 150, 0));
+
+    pixels->show();   // Send the updated pixel colors to the hardware.
+
+    delay(DELAYVAL); // Pause before next pass through loop
+  }
+}

Adafruit_NeoPixel/examples/strandtest/strandtest.ino

@@ -0,0 +1,147 @@
+// A basic everyday NeoPixel strip test program.
+
+// NEOPIXEL BEST PRACTICES for most reliable operation:
+// - Add 1000 uF CAPACITOR between NeoPixel strip's + and - connections.
+// - MINIMIZE WIRING LENGTH between microcontroller board and first pixel.
+// - NeoPixel strip's DATA-IN should pass through a 300-500 OHM RESISTOR.
+// - AVOID connecting NeoPixels on a LIVE CIRCUIT. If you must, ALWAYS
+//   connect GROUND (-) first, then +, then data.
+// - When using a 3.3V microcontroller with a 5V-powered NeoPixel strip,
+//   a LOGIC-LEVEL CONVERTER on the data line is STRONGLY RECOMMENDED.
+// (Skipping these may work OK on your workbench but can fail in the field)
+
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+ #include <avr/power.h> // Required for 16 MHz Adafruit Trinket
+#endif
+
+// Which pin on the Arduino is connected to the NeoPixels?
+// On a Trinket or Gemma we suggest changing this to 1:
+#define LED_PIN    6
+
+// How many NeoPixels are attached to the Arduino?
+#define LED_COUNT 60
+
+// Declare our NeoPixel strip object:
+Adafruit_NeoPixel strip(LED_COUNT, LED_PIN, NEO_GRB + NEO_KHZ800);
+// Argument 1 = Number of pixels in NeoPixel strip
+// Argument 2 = Arduino pin number (most are valid)
+// Argument 3 = Pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+
+
+// setup() function -- runs once at startup --------------------------------
+
+void setup() {
+  // These lines are specifically to support the Adafruit Trinket 5V 16 MHz.
+  // Any other board, you can remove this part (but no harm leaving it):
+#if defined(__AVR_ATtiny85__) && (F_CPU == 16000000)
+  clock_prescale_set(clock_div_1);
+#endif
+  // END of Trinket-specific code.
+
+  strip.begin();           // INITIALIZE NeoPixel strip object (REQUIRED)
+  strip.show();            // Turn OFF all pixels ASAP
+  strip.setBrightness(50); // Set BRIGHTNESS to about 1/5 (max = 255)
+}
+
+
+// loop() function -- runs repeatedly as long as board is on ---------------
+
+void loop() {
+  // Fill along the length of the strip in various colors...
+  colorWipe(strip.Color(255,   0,   0), 50); // Red
+  colorWipe(strip.Color(  0, 255,   0), 50); // Green
+  colorWipe(strip.Color(  0,   0, 255), 50); // Blue
+
+  // Do a theater marquee effect in various colors...
+  theaterChase(strip.Color(127, 127, 127), 50); // White, half brightness
+  theaterChase(strip.Color(127,   0,   0), 50); // Red, half brightness
+  theaterChase(strip.Color(  0,   0, 127), 50); // Blue, half brightness
+
+  rainbow(10);             // Flowing rainbow cycle along the whole strip
+  theaterChaseRainbow(50); // Rainbow-enhanced theaterChase variant
+}
+
+
+// Some functions of our own for creating animated effects -----------------
+
+// Fill strip pixels one after another with a color. Strip is NOT cleared
+// first; anything there will be covered pixel by pixel. Pass in color
+// (as a single 'packed' 32-bit value, which you can get by calling
+// strip.Color(red, green, blue) as shown in the loop() function above),
+// and a delay time (in milliseconds) between pixels.
+void colorWipe(uint32_t color, int wait) {
+  for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+    strip.setPixelColor(i, color);         //  Set pixel's color (in RAM)
+    strip.show();                          //  Update strip to match
+    delay(wait);                           //  Pause for a moment
+  }
+}
+
+// Theater-marquee-style chasing lights. Pass in a color (32-bit value,
+// a la strip.Color(r,g,b) as mentioned above), and a delay time (in ms)
+// between frames.
+void theaterChase(uint32_t color, int wait) {
+  for(int a=0; a<10; a++) {  // Repeat 10 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in steps of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show(); // Update strip with new contents
+      delay(wait);  // Pause for a moment
+    }
+  }
+}
+
+// Rainbow cycle along whole strip. Pass delay time (in ms) between frames.
+void rainbow(int wait) {
+  // Hue of first pixel runs 5 complete loops through the color wheel.
+  // Color wheel has a range of 65536 but it's OK if we roll over, so
+  // just count from 0 to 5*65536. Adding 256 to firstPixelHue each time
+  // means we'll make 5*65536/256 = 1280 passes through this outer loop:
+  for(long firstPixelHue = 0; firstPixelHue < 5*65536; firstPixelHue += 256) {
+    for(int i=0; i<strip.numPixels(); i++) { // For each pixel in strip...
+      // Offset pixel hue by an amount to make one full revolution of the
+      // color wheel (range of 65536) along the length of the strip
+      // (strip.numPixels() steps):
+      int pixelHue = firstPixelHue + (i * 65536L / strip.numPixels());
+      // strip.ColorHSV() can take 1 or 3 arguments: a hue (0 to 65535) or
+      // optionally add saturation and value (brightness) (each 0 to 255).
+      // Here we're using just the single-argument hue variant. The result
+      // is passed through strip.gamma32() to provide 'truer' colors
+      // before assigning to each pixel:
+      strip.setPixelColor(i, strip.gamma32(strip.ColorHSV(pixelHue)));
+    }
+    strip.show(); // Update strip with new contents
+    delay(wait);  // Pause for a moment
+  }
+}
+
+// Rainbow-enhanced theater marquee. Pass delay time (in ms) between frames.
+void theaterChaseRainbow(int wait) {
+  int firstPixelHue = 0;     // First pixel starts at red (hue 0)
+  for(int a=0; a<30; a++) {  // Repeat 30 times...
+    for(int b=0; b<3; b++) { //  'b' counts from 0 to 2...
+      strip.clear();         //   Set all pixels in RAM to 0 (off)
+      // 'c' counts up from 'b' to end of strip in increments of 3...
+      for(int c=b; c<strip.numPixels(); c += 3) {
+        // hue of pixel 'c' is offset by an amount to make one full
+        // revolution of the color wheel (range 65536) along the length
+        // of the strip (strip.numPixels() steps):
+        int      hue   = firstPixelHue + c * 65536L / strip.numPixels();
+        uint32_t color = strip.gamma32(strip.ColorHSV(hue)); // hue -> RGB
+        strip.setPixelColor(c, color); // Set pixel 'c' to value 'color'
+      }
+      strip.show();                // Update strip with new contents
+      delay(wait);                 // Pause for a moment
+      firstPixelHue += 65536 / 90; // One cycle of color wheel over 90 frames
+    }
+  }
+}

Adafruit_NeoPixel/examples/strandtest_wheel/strandtest_wheel.ino

@@ -0,0 +1,134 @@
+#include <Adafruit_NeoPixel.h>
+#ifdef __AVR__
+  #include <avr/power.h>
+#endif
+
+#define PIN 6
+
+// Parameter 1 = number of pixels in strip
+// Parameter 2 = Arduino pin number (most are valid)
+// Parameter 3 = pixel type flags, add together as needed:
+//   NEO_KHZ800  800 KHz bitstream (most NeoPixel products w/WS2812 LEDs)
+//   NEO_KHZ400  400 KHz (classic 'v1' (not v2) FLORA pixels, WS2811 drivers)
+//   NEO_GRB     Pixels are wired for GRB bitstream (most NeoPixel products)
+//   NEO_RGB     Pixels are wired for RGB bitstream (v1 FLORA pixels, not v2)
+//   NEO_RGBW    Pixels are wired for RGBW bitstream (NeoPixel RGBW products)
+Adafruit_NeoPixel strip = Adafruit_NeoPixel(60, PIN, NEO_GRB + NEO_KHZ800);
+
+// IMPORTANT: To reduce NeoPixel burnout risk, add 1000 uF capacitor across
+// pixel power leads, add 300 - 500 Ohm resistor on first pixel's data input
+// and minimize distance between Arduino and first pixel.  Avoid connecting
+// on a live circuit...if you must, connect GND first.
+
+void setup() {
+  // This is for Trinket 5V 16MHz, you can remove these three lines if you are not using a Trinket
+  #if defined (__AVR_ATtiny85__)
+    if (F_CPU == 16000000) clock_prescale_set(clock_div_1);
+  #endif
+  // End of trinket special code
+
+  strip.begin();
+  strip.setBrightness(50);
+  strip.show(); // Initialize all pixels to 'off'
+}
+
+void loop() {
+  // Some example procedures showing how to display to the pixels:
+  colorWipe(strip.Color(255, 0, 0), 50); // Red
+  colorWipe(strip.Color(0, 255, 0), 50); // Green
+  colorWipe(strip.Color(0, 0, 255), 50); // Blue
+//colorWipe(strip.Color(0, 0, 0, 255), 50); // White RGBW
+  // Send a theater pixel chase in...
+  theaterChase(strip.Color(127, 127, 127), 50); // White
+  theaterChase(strip.Color(127, 0, 0), 50); // Red
+  theaterChase(strip.Color(0, 0, 127), 50); // Blue
+
+  rainbow(20);
+  rainbowCycle(20);
+  theaterChaseRainbow(50);
+}
+
+// Fill the dots one after the other with a color
+void colorWipe(uint32_t c, uint8_t wait) {
+  for(uint16_t i=0; i<strip.numPixels(); i++) {
+    strip.setPixelColor(i, c);
+    strip.show();
+    delay(wait);
+  }
+}
+
+void rainbow(uint8_t wait) {
+  uint16_t i, j;
+
+  for(j=0; j<256; j++) {
+    for(i=0; i<strip.numPixels(); i++) {
+      strip.setPixelColor(i, Wheel((i+j) & 255));
+    }
+    strip.show();
+    delay(wait);
+  }
+}
+
+// Slightly different, this makes the rainbow equally distributed throughout
+void rainbowCycle(uint8_t wait) {
+  uint16_t i, j;
+
+  for(j=0; j<256*5; j++) { // 5 cycles of all colors on wheel
+    for(i=0; i< strip.numPixels(); i++) {
+      strip.setPixelColor(i, Wheel(((i * 256 / strip.numPixels()) + j) & 255));
+    }
+    strip.show();
+    delay(wait);
+  }
+}
+
+//Theatre-style crawling lights.
+void theaterChase(uint32_t c, uint8_t wait) {
+  for (int j=0; j<10; j++) {  //do 10 cycles of chasing
+    for (int q=0; q < 3; q++) {
+      for (uint16_t i=0; i < strip.numPixels(); i=i+3) {
+        strip.setPixelColor(i+q, c);    //turn every third pixel on
+      }
+      strip.show();
+
+      delay(wait);
+
+      for (uint16_t i=0; i < strip.numPixels(); i=i+3) {
+        strip.setPixelColor(i+q, 0);        //turn every third pixel off
+      }
+    }
+  }
+}
+
+//Theatre-style crawling lights with rainbow effect
+void theaterChaseRainbow(uint8_t wait) {
+  for (int j=0; j < 256; j++) {     // cycle all 256 colors in the wheel
+    for (int q=0; q < 3; q++) {
+      for (uint16_t i=0; i < strip.numPixels(); i=i+3) {
+        strip.setPixelColor(i+q, Wheel( (i+j) % 255));    //turn every third pixel on
+      }
+      strip.show();
+
+      delay(wait);
+
+      for (uint16_t i=0; i < strip.numPixels(); i=i+3) {
+        strip.setPixelColor(i+q, 0);        //turn every third pixel off
+      }
+    }
+  }
+}
+
+// Input a value 0 to 255 to get a color value.
+// The colours are a transition r - g - b - back to r.
+uint32_t Wheel(byte WheelPos) {
+  WheelPos = 255 - WheelPos;
+  if(WheelPos < 85) {
+    return strip.Color(255 - WheelPos * 3, 0, WheelPos * 3);
+  }
+  if(WheelPos < 170) {
+    WheelPos -= 85;
+    return strip.Color(0, WheelPos * 3, 255 - WheelPos * 3);
+  }
+  WheelPos -= 170;
+  return strip.Color(WheelPos * 3, 255 - WheelPos * 3, 0);
+}

Adafruit_NeoPixel/kendyte_k210.c

@@ -0,0 +1,74 @@
+// This is a mash-up of the Due show() code + insights from Michael Miller's
+// ESP8266 work for the NeoPixelBus library: github.com/Makuna/NeoPixelBus
+// Needs to be a separate .c file to enforce ICACHE_RAM_ATTR execution.
+#if defined(K210) 
+#define KENDRYTE_K210 1
+#endif
+
+#if defined(KENDRYTE_K210)
+
+#include <Arduino.h>
+#include "sysctl.h"
+
+void  k210Show(
+    uint8_t pin, uint8_t *pixels, uint32_t numBytes, boolean is800KHz)
+{
+
+#define CYCLES_800_T0H (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 2500000) // 0.4us
+#define CYCLES_800_T1H (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 1250000) // 0.8us
+#define CYCLES_800 (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 800000)      // 1.25us per bit
+#define CYCLES_400_T0H (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 2000000) // 0.5uS
+#define CYCLES_400_T1H (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 833333)  // 1.2us
+#define CYCLES_400 (sysctl_clock_get_freq(SYSCTL_CLOCK_CPU) / 400000)      // 2.5us per bit
+
+    uint8_t *p, *end, pix, mask;
+    uint32_t t, time0, time1, period, c, startTime;
+
+    p = pixels;
+    end = p + numBytes;
+    pix = *p++;
+    mask = 0x80;
+    startTime = 0;
+
+#ifdef NEO_KHZ400
+    if (is800KHz)
+    {
+#endif
+        time0 = CYCLES_800_T0H;
+        time1 = CYCLES_800_T1H;
+        period = CYCLES_800;
+#ifdef NEO_KHZ400
+    }
+    else
+    { // 400 KHz bitstream
+        time0 = CYCLES_400_T0H;
+        time1 = CYCLES_400_T1H;
+        period = CYCLES_400;
+    }
+#endif
+
+    for (t = time0;; t = time0)
+    {
+        if (pix & mask)
+            t = time1; // Bit high duration
+        while (((c = read_cycle()) - startTime) < period)
+            ; // Wait for bit start
+        digitalWrite(pin, HIGH);
+        startTime = c; // Save start time
+        while (((c = read_cycle()) - startTime) < t)
+            ; // Wait high duration
+        digitalWrite(pin, LOW);
+
+        if (!(mask >>= 1))
+        { // Next bit/byte
+            if (p >= end)
+                break;
+            pix = *p++;
+            mask = 0x80;
+        }
+    }
+    while ((read_cycle() - startTime) < period)
+        ; // Wait for last bit
+}
+
+#endif // KENDRYTE_K210

Adafruit_NeoPixel/keywords.txt

@@ -0,0 +1,72 @@
+#######################################
+# Syntax Coloring Map For Adafruit_NeoPixel
+#######################################
+# Class
+#######################################
+
+Adafruit_NeoPixel	KEYWORD1
+
+#######################################
+# Methods and Functions
+#######################################	
+
+begin			KEYWORD2
+show			KEYWORD2
+setPin			KEYWORD2
+setPixelColor		KEYWORD2
+fill			KEYWORD2
+setBrightness		KEYWORD2
+clear			KEYWORD2
+updateLength		KEYWORD2
+updateType		KEYWORD2
+canShow			KEYWORD2
+getPixels		KEYWORD2
+getBrightness		KEYWORD2
+getPin			KEYWORD2
+numPixels		KEYWORD2
+getPixelColor		KEYWORD2
+sine8			KEYWORD2
+gamma8			KEYWORD2
+Color			KEYWORD2
+ColorHSV		KEYWORD2
+gamma32			KEYWORD2
+
+#######################################
+# Constants
+#######################################
+
+NEO_COLMASK		LITERAL1
+NEO_SPDMASK		LITERAL1
+NEO_KHZ800		LITERAL1
+NEO_KHZ400		LITERAL1
+NEO_RGB			LITERAL1
+NEO_RBG			LITERAL1
+NEO_GRB			LITERAL1
+NEO_GBR			LITERAL1
+NEO_BRG			LITERAL1
+NEO_BGR			LITERAL1
+NEO_WRGB		LITERAL1
+NEO_WRBG		LITERAL1
+NEO_WGRB		LITERAL1
+NEO_WGBR		LITERAL1
+NEO_WBRG		LITERAL1
+NEO_WBGR		LITERAL1
+NEO_RWGB		LITERAL1
+NEO_RWBG		LITERAL1
+NEO_RGWB		LITERAL1
+NEO_RGBW		LITERAL1
+NEO_RBWG		LITERAL1
+NEO_RBGW		LITERAL1
+NEO_GWRB		LITERAL1
+NEO_GWBR		LITERAL1
+NEO_GRWB		LITERAL1
+NEO_GRBW		LITERAL1
+NEO_GBWR		LITERAL1
+NEO_GBRW		LITERAL1
+NEO_BWRG		LITERAL1
+NEO_BWGR		LITERAL1
+NEO_BRWG		LITERAL1
+NEO_BRGW		LITERAL1
+NEO_BGWR		LITERAL1
+NEO_BGRW		LITERAL1
+

Adafruit_NeoPixel/library.properties

@@ -0,0 +1,9 @@
+name=Adafruit NeoPixel
+version=1.8.7
+author=Adafruit
+maintainer=Adafruit <info@adafruit.com>
+sentence=Arduino library for controlling single-wire-based LED pixels and strip.
+paragraph=Arduino library for controlling single-wire-based LED pixels and strip.
+category=Display
+url=https://github.com/adafruit/Adafruit_NeoPixel
+architectures=*

Adafruit_NeoPixel/rp2040.c

@@ -0,0 +1,53 @@
+//  This sketch is based on the SDK example here:
+//  https://github.com/raspberrypi/pico-examples/tree/master/pio/ws2812
+
+/**
+   Copyright (c) 2020 Raspberry Pi (Trading) Ltd.
+
+   SPDX-License-Identifier: BSD-3-Clause
+*/
+
+#if defined(ARDUINO_ARCH_RP2040)
+
+#include <stdlib.h>
+#include "hardware/pio.h"
+#include "hardware/clocks.h"
+
+#include "rp2040_pio.h"
+
+void rp2040Init(uint8_t pin, bool is800KHz)
+{
+    // todo get free sm
+    PIO pio = pio0;
+    int sm = 0;
+    uint offset = pio_add_program(pio, &ws2812_program);
+
+    if (is800KHz)
+    {
+        // 800kHz, 8 bit transfers
+        ws2812_program_init(pio, sm, offset, pin, 800000, 8);
+    }
+    else
+    {
+        // 400kHz, 8 bit transfers
+        ws2812_program_init(pio, sm, offset, pin, 400000, 8);
+    }
+}
+ 
+void  rp2040Show(uint8_t pin, uint8_t *pixels, uint32_t numBytes, bool is800KHz)
+{
+    static bool init = true;
+    
+    if (init)
+    {
+        // On first pass through initialise the PIO
+        rp2040Init(pin, is800KHz);
+        init = false;
+    }
+
+    while(numBytes--)
+        // Bits for transmission must be shifted to top 8 bits
+        pio_sm_put_blocking(pio0, 0, ((uint32_t)*pixels++)<< 24);
+}
+
+#endif // KENDRYTE_K210

Adafruit_NeoPixel/rp2040_pio.h

@@ -0,0 +1,61 @@
+// -------------------------------------------------- //
+// This file is autogenerated by pioasm; do not edit! //
+// -------------------------------------------------- //
+
+// Unless you know what you are doing... 
+// Lines 47 and 52 have been edited to set transmit bit count
+
+#if !PICO_NO_HARDWARE
+#include "hardware/pio.h"
+#endif
+
+// ------ //
+// ws2812 //
+// ------ //
+
+#define ws2812_wrap_target 0
+#define ws2812_wrap 3
+
+#define ws2812_T1 2
+#define ws2812_T2 5
+#define ws2812_T3 3
+
+static const uint16_t ws2812_program_instructions[] = {
+            //     .wrap_target
+    0x6221, //  0: out    x, 1            side 0 [2] 
+    0x1123, //  1: jmp    !x, 3           side 1 [1] 
+    0x1400, //  2: jmp    0               side 1 [4] 
+    0xa442, //  3: nop                    side 0 [4] 
+            //     .wrap
+};
+
+#if !PICO_NO_HARDWARE
+static const struct pio_program ws2812_program = {
+    .instructions = ws2812_program_instructions,
+    .length = 4,
+    .origin = -1,
+};
+
+static inline pio_sm_config ws2812_program_get_default_config(uint offset) {
+    pio_sm_config c = pio_get_default_sm_config();
+    sm_config_set_wrap(&c, offset + ws2812_wrap_target, offset + ws2812_wrap);
+    sm_config_set_sideset(&c, 1, false, false);
+    return c;
+}
+
+#include "hardware/clocks.h"
+static inline void ws2812_program_init(PIO pio, uint sm, uint offset, uint pin, float freq, uint bits) {
+    pio_gpio_init(pio, pin);
+    pio_sm_set_consecutive_pindirs(pio, sm, pin, 1, true);
+    pio_sm_config c = ws2812_program_get_default_config(offset);
+    sm_config_set_sideset_pins(&c, pin);
+    sm_config_set_out_shift(&c, false, true, bits);    // <----<<< Length changed to "bits"
+    sm_config_set_fifo_join(&c, PIO_FIFO_JOIN_TX);
+    int cycles_per_bit = ws2812_T1 + ws2812_T2 + ws2812_T3;
+    float div = clock_get_hz(clk_sys) / (freq * cycles_per_bit);
+    sm_config_set_clkdiv(&c, div);
+    pio_sm_init(pio, sm, offset, &c);
+    pio_sm_set_enabled(pio, sm, true);
+}
+
+#endif

CMakeLists.txt

@@ -0,0 +1,30 @@
+cmake_minimum_required(VERSION 3.12)
+
+include(pico_sdk_import.cmake)
+
+project(my_project)
+
+pico_sdk_init()
+
+
+add_executable(adc_fft adc_fft.c)
+add_library(kiss_fftr kiss_fftr.c)
+add_library(kiss_fft kiss_fft.c)
+
+target_link_libraries(kiss_fftr kiss_fft)
+
+pico_enable_stdio_usb(adc_fft 1)
+pico_enable_stdio_uart(adc_fft 1)
+
+pico_add_extra_outputs(adc_fft)
+
+# generate the header file into the source tree as it is included in the RP2040 datasheet
+pico_generate_pio_header(adc_fft ${CMAKE_CURRENT_LIST_DIR}/ws2812.pio OUTPUT_DIR ${CMAKE_CURRENT_LIST_DIR}/generated)
+
+target_link_libraries(adc_fft
+	pico_stdlib
+	hardware_adc
+	hardware_dma
+	kiss_fftr
+		hardware_pio
+	)

adc_fft.c

@@ -0,0 +1,167 @@
+// Sample from the ADC continuously at a particular sample rate
+// and then compute an FFT over the data
+//
+// much of this code is from pico-examples/adc/dma_capture/dma_capture.c
+// the rest is written by Alex Wulff (www.AlexWulff.com)
+
+#include <stdio.h>
+#include <math.h>
+
+#include "pico/stdlib.h"
+#include "hardware/adc.h"
+#include "hardware/dma.h"
+#include "kiss_fftr.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "pico/stdlib.h"
+#include "hardware/pio.h"
+#include "hardware/clocks.h"
+#include "ws2812.pio.h"
+
+#define IS_RGBW false
+#define NUM_PIXELS 150
+
+#ifdef PICO_DEFAULT_WS2812_PIN
+#define WS2812_PIN PICO_DEFAULT_WS2812_PIN
+#else
+// default to pin 2 if the board doesn't have a default WS2812 pin defined
+#define WS2812_PIN 6
+#endif
+
+static inline void put_pixel(uint32_t pixel_grb) {
+    pio_sm_put_blocking(pio0, 0, pixel_grb << 8u);
+}
+
+// set this to determine sample rate
+// 0     = 500,000 Hz
+// 960   = 50,000 Hz
+// 9600  = 5,000 Hz
+#define CLOCK_DIV 960
+#define FSAMP 50000
+
+// Channel 0 is GPIO26
+#define CAPTURE_CHANNEL 0
+#define LED_PIN 25
+
+// BE CAREFUL: anything over about 9000 here will cause things
+// to silently break. The code will compile and upload, but due
+// to memory issues nothing will work properly
+#define NSAMP 4096
+
+// globals
+dma_channel_config cfg;
+uint dma_chan;
+float freqs[NSAMP];
+
+void setup();
+void sample(uint8_t *capture_buf);
+
+int main() {
+
+
+  uint8_t cap_buf[NSAMP];
+  kiss_fft_scalar fft_in[NSAMP]; // kiss_fft_scalar is a float
+  kiss_fft_cpx fft_out[NSAMP];
+  kiss_fftr_cfg cfg = kiss_fftr_alloc(NSAMP,false,0,0);
+  
+  // setup ports and outputs
+  setup();
+
+  while (1) {
+    // get NSAMP samples at FSAMP
+    sample(cap_buf);
+    // fill fourier transform input while subtracting DC component
+    uint64_t sum = 0;
+    for (int i=0;i<NSAMP;i++) {sum+=cap_buf[i];}
+    float avg = (float)sum/NSAMP;
+    for (int i=0;i<NSAMP;i++) {fft_in[i]=(float)cap_buf[i]-avg;}
+
+    // compute fast fourier transform
+    kiss_fftr(cfg , fft_in, fft_out);
+    
+    // compute power and calculate max freq component
+    float max_power = 0;
+    int max_idx = 0;
+    // any frequency bin over NSAMP/2 is aliased (nyquist sampling theorum)
+    for (int i = 0; i < NSAMP/2; i++) {
+      float power = fft_out[i].r*fft_out[i].r+fft_out[i].i*fft_out[i].i;
+      if (power>max_power) {
+	max_power=power;
+	max_idx = i;
+      }
+    }
+
+    float max_freq = freqs[max_idx];
+    printf("Greatest Frequency Component: %0.1f Hz\n",max_freq);
+    put_pixel(0x00ffff);
+    put_pixel(0xff00ff);
+  }
+
+  // should never get here
+  kiss_fft_free(cfg);
+}
+
+void sample(uint8_t *capture_buf) {
+  adc_fifo_drain();
+  adc_run(false);
+      
+  dma_channel_configure(dma_chan, &cfg,
+			capture_buf,    // dst
+			&adc_hw->fifo,  // src
+			NSAMP,          // transfer count
+			true            // start immediately
+			);
+
+  gpio_put(LED_PIN, 1);
+  adc_run(true);
+  dma_channel_wait_for_finish_blocking(dma_chan);
+  gpio_put(LED_PIN, 0);
+}
+
+void setup() {
+  stdio_init_all();
+
+  // todo get free sm
+  PIO pio = pio0;
+  int sm = 0;
+  uint offset = pio_add_program(pio, &ws2812_program);
+  ws2812_program_init(pio, sm, offset, WS2812_PIN, 800000, IS_RGBW);
+
+  //gpio_init(LED_PIN);
+  //gpio_set_dir(LED_PIN, GPIO_OUT);
+
+  adc_gpio_init(26 + CAPTURE_CHANNEL);
+
+  adc_init();
+  adc_select_input(CAPTURE_CHANNEL);
+  adc_fifo_setup(
+		 true,    // Write each completed conversion to the sample FIFO
+		 true,    // Enable DMA data request (DREQ)
+		 1,       // DREQ (and IRQ) asserted when at least 1 sample present
+		 false,   // We won't see the ERR bit because of 8 bit reads; disable.
+		 true     // Shift each sample to 8 bits when pushing to FIFO
+		 );
+
+  // set sample rate
+  adc_set_clkdiv(CLOCK_DIV);
+
+  sleep_ms(1000);
+  // Set up the DMA to start transferring data as soon as it appears in FIFO
+  uint dma_chan = dma_claim_unused_channel(true);
+  cfg = dma_channel_get_default_config(dma_chan);
+
+  // Reading from constant address, writing to incrementing byte addresses
+  channel_config_set_transfer_data_size(&cfg, DMA_SIZE_8);
+  channel_config_set_read_increment(&cfg, false);
+  channel_config_set_write_increment(&cfg, true);
+
+  // Pace transfers based on availability of ADC samples
+  channel_config_set_dreq(&cfg, DREQ_ADC);
+
+  // calculate frequencies of each bin
+  float f_max = FSAMP;
+  float f_res = f_max / NSAMP;
+  for (int i = 0; i < NSAMP; i++) {freqs[i] = f_res*i;}
+}

kiss_fft.c

@@ -0,0 +1,408 @@
+/*
+Copyright (c) 2003-2010, Mark Borgerding
+
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+    * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+    * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+
+#include "_kiss_fft_guts.h"
+/* The guts header contains all the multiplication and addition macros that are defined for
+ fixed or floating point complex numbers.  It also delares the kf_ internal functions.
+ */
+
+static void kf_bfly2(
+        kiss_fft_cpx * Fout,
+        const size_t fstride,
+        const kiss_fft_cfg st,
+        int m
+        )
+{
+    kiss_fft_cpx * Fout2;
+    kiss_fft_cpx * tw1 = st->twiddles;
+    kiss_fft_cpx t;
+    Fout2 = Fout + m;
+    do{
+        C_FIXDIV(*Fout,2); C_FIXDIV(*Fout2,2);
+
+        C_MUL (t,  *Fout2 , *tw1);
+        tw1 += fstride;
+        C_SUB( *Fout2 ,  *Fout , t );
+        C_ADDTO( *Fout ,  t );
+        ++Fout2;
+        ++Fout;
+    }while (--m);
+}
+
+static void kf_bfly4(
+        kiss_fft_cpx * Fout,
+        const size_t fstride,
+        const kiss_fft_cfg st,
+        const size_t m
+        )
+{
+    kiss_fft_cpx *tw1,*tw2,*tw3;
+    kiss_fft_cpx scratch[6];
+    size_t k=m;
+    const size_t m2=2*m;
+    const size_t m3=3*m;
+
+
+    tw3 = tw2 = tw1 = st->twiddles;
+
+    do {
+        C_FIXDIV(*Fout,4); C_FIXDIV(Fout[m],4); C_FIXDIV(Fout[m2],4); C_FIXDIV(Fout[m3],4);
+
+        C_MUL(scratch[0],Fout[m] , *tw1 );
+        C_MUL(scratch[1],Fout[m2] , *tw2 );
+        C_MUL(scratch[2],Fout[m3] , *tw3 );
+
+        C_SUB( scratch[5] , *Fout, scratch[1] );
+        C_ADDTO(*Fout, scratch[1]);
+        C_ADD( scratch[3] , scratch[0] , scratch[2] );
+        C_SUB( scratch[4] , scratch[0] , scratch[2] );
+        C_SUB( Fout[m2], *Fout, scratch[3] );
+        tw1 += fstride;
+        tw2 += fstride*2;
+        tw3 += fstride*3;
+        C_ADDTO( *Fout , scratch[3] );
+
+        if(st->inverse) {
+            Fout[m].r = scratch[5].r - scratch[4].i;
+            Fout[m].i = scratch[5].i + scratch[4].r;
+            Fout[m3].r = scratch[5].r + scratch[4].i;
+            Fout[m3].i = scratch[5].i - scratch[4].r;
+        }else{
+            Fout[m].r = scratch[5].r + scratch[4].i;
+            Fout[m].i = scratch[5].i - scratch[4].r;
+            Fout[m3].r = scratch[5].r - scratch[4].i;
+            Fout[m3].i = scratch[5].i + scratch[4].r;
+        }
+        ++Fout;
+    }while(--k);
+}
+
+static void kf_bfly3(
+         kiss_fft_cpx * Fout,
+         const size_t fstride,
+         const kiss_fft_cfg st,
+         size_t m
+         )
+{
+     size_t k=m;
+     const size_t m2 = 2*m;
+     kiss_fft_cpx *tw1,*tw2;
+     kiss_fft_cpx scratch[5];
+     kiss_fft_cpx epi3;
+     epi3 = st->twiddles[fstride*m];
+
+     tw1=tw2=st->twiddles;
+
+     do{
+         C_FIXDIV(*Fout,3); C_FIXDIV(Fout[m],3); C_FIXDIV(Fout[m2],3);
+
+         C_MUL(scratch[1],Fout[m] , *tw1);
+         C_MUL(scratch[2],Fout[m2] , *tw2);
+
+         C_ADD(scratch[3],scratch[1],scratch[2]);
+         C_SUB(scratch[0],scratch[1],scratch[2]);
+         tw1 += fstride;
+         tw2 += fstride*2;
+
+         Fout[m].r = Fout->r - HALF_OF(scratch[3].r);
+         Fout[m].i = Fout->i - HALF_OF(scratch[3].i);
+
+         C_MULBYSCALAR( scratch[0] , epi3.i );
+
+         C_ADDTO(*Fout,scratch[3]);
+
+         Fout[m2].r = Fout[m].r + scratch[0].i;
+         Fout[m2].i = Fout[m].i - scratch[0].r;
+
+         Fout[m].r -= scratch[0].i;
+         Fout[m].i += scratch[0].r;
+
+         ++Fout;
+     }while(--k);
+}
+
+static void kf_bfly5(
+        kiss_fft_cpx * Fout,
+        const size_t fstride,
+        const kiss_fft_cfg st,
+        int m
+        )
+{
+    kiss_fft_cpx *Fout0,*Fout1,*Fout2,*Fout3,*Fout4;
+    int u;
+    kiss_fft_cpx scratch[13];
+    kiss_fft_cpx * twiddles = st->twiddles;
+    kiss_fft_cpx *tw;
+    kiss_fft_cpx ya,yb;
+    ya = twiddles[fstride*m];
+    yb = twiddles[fstride*2*m];
+
+    Fout0=Fout;
+    Fout1=Fout0+m;
+    Fout2=Fout0+2*m;
+    Fout3=Fout0+3*m;
+    Fout4=Fout0+4*m;
+
+    tw=st->twiddles;
+    for ( u=0; u<m; ++u ) {
+        C_FIXDIV( *Fout0,5); C_FIXDIV( *Fout1,5); C_FIXDIV( *Fout2,5); C_FIXDIV( *Fout3,5); C_FIXDIV( *Fout4,5);
+        scratch[0] = *Fout0;
+
+        C_MUL(scratch[1] ,*Fout1, tw[u*fstride]);
+        C_MUL(scratch[2] ,*Fout2, tw[2*u*fstride]);
+        C_MUL(scratch[3] ,*Fout3, tw[3*u*fstride]);
+        C_MUL(scratch[4] ,*Fout4, tw[4*u*fstride]);
+
+        C_ADD( scratch[7],scratch[1],scratch[4]);
+        C_SUB( scratch[10],scratch[1],scratch[4]);
+        C_ADD( scratch[8],scratch[2],scratch[3]);
+        C_SUB( scratch[9],scratch[2],scratch[3]);
+
+        Fout0->r += scratch[7].r + scratch[8].r;
+        Fout0->i += scratch[7].i + scratch[8].i;
+
+        scratch[5].r = scratch[0].r + S_MUL(scratch[7].r,ya.r) + S_MUL(scratch[8].r,yb.r);
+        scratch[5].i = scratch[0].i + S_MUL(scratch[7].i,ya.r) + S_MUL(scratch[8].i,yb.r);
+
+        scratch[6].r =  S_MUL(scratch[10].i,ya.i) + S_MUL(scratch[9].i,yb.i);
+        scratch[6].i = -S_MUL(scratch[10].r,ya.i) - S_MUL(scratch[9].r,yb.i);
+
+        C_SUB(*Fout1,scratch[5],scratch[6]);
+        C_ADD(*Fout4,scratch[5],scratch[6]);
+
+        scratch[11].r = scratch[0].r + S_MUL(scratch[7].r,yb.r) + S_MUL(scratch[8].r,ya.r);
+        scratch[11].i = scratch[0].i + S_MUL(scratch[7].i,yb.r) + S_MUL(scratch[8].i,ya.r);
+        scratch[12].r = - S_MUL(scratch[10].i,yb.i) + S_MUL(scratch[9].i,ya.i);
+        scratch[12].i = S_MUL(scratch[10].r,yb.i) - S_MUL(scratch[9].r,ya.i);
+
+        C_ADD(*Fout2,scratch[11],scratch[12]);
+        C_SUB(*Fout3,scratch[11],scratch[12]);
+
+        ++Fout0;++Fout1;++Fout2;++Fout3;++Fout4;
+    }
+}
+
+/* perform the butterfly for one stage of a mixed radix FFT */
+static void kf_bfly_generic(
+        kiss_fft_cpx * Fout,
+        const size_t fstride,
+        const kiss_fft_cfg st,
+        int m,
+        int p
+        )
+{
+    int u,k,q1,q;
+    kiss_fft_cpx * twiddles = st->twiddles;
+    kiss_fft_cpx t;
+    int Norig = st->nfft;
+
+    kiss_fft_cpx * scratch = (kiss_fft_cpx*)KISS_FFT_TMP_ALLOC(sizeof(kiss_fft_cpx)*p);
+
+    for ( u=0; u<m; ++u ) {
+        k=u;
+        for ( q1=0 ; q1<p ; ++q1 ) {
+            scratch[q1] = Fout[ k  ];
+            C_FIXDIV(scratch[q1],p);
+            k += m;
+        }
+
+        k=u;
+        for ( q1=0 ; q1<p ; ++q1 ) {
+            int twidx=0;
+            Fout[ k ] = scratch[0];
+            for (q=1;q<p;++q ) {
+                twidx += fstride * k;
+                if (twidx>=Norig) twidx-=Norig;
+                C_MUL(t,scratch[q] , twiddles[twidx] );
+                C_ADDTO( Fout[ k ] ,t);
+            }
+            k += m;
+        }
+    }
+    KISS_FFT_TMP_FREE(scratch);
+}
+
+static
+void kf_work(
+        kiss_fft_cpx * Fout,
+        const kiss_fft_cpx * f,
+        const size_t fstride,
+        int in_stride,
+        int * factors,
+        const kiss_fft_cfg st
+        )
+{
+    kiss_fft_cpx * Fout_beg=Fout;
+    const int p=*factors++; /* the radix  */
+    const int m=*factors++; /* stage's fft length/p */
+    const kiss_fft_cpx * Fout_end = Fout + p*m;
+
+#ifdef _OPENMP
+    // use openmp extensions at the 
+    // top-level (not recursive)
+    if (fstride==1 && p<=5)
+    {
+        int k;
+
+        // execute the p different work units in different threads
+#       pragma omp parallel for
+        for (k=0;k<p;++k) 
+            kf_work( Fout +k*m, f+ fstride*in_stride*k,fstride*p,in_stride,factors,st);
+        // all threads have joined by this point
+
+        switch (p) {
+            case 2: kf_bfly2(Fout,fstride,st,m); break;
+            case 3: kf_bfly3(Fout,fstride,st,m); break; 
+            case 4: kf_bfly4(Fout,fstride,st,m); break;
+            case 5: kf_bfly5(Fout,fstride,st,m); break; 
+            default: kf_bfly_generic(Fout,fstride,st,m,p); break;
+        }
+        return;
+    }
+#endif
+
+    if (m==1) {
+        do{
+            *Fout = *f;
+            f += fstride*in_stride;
+        }while(++Fout != Fout_end );
+    }else{
+        do{
+            // recursive call:
+            // DFT of size m*p performed by doing
+            // p instances of smaller DFTs of size m, 
+            // each one takes a decimated version of the input
+            kf_work( Fout , f, fstride*p, in_stride, factors,st);
+            f += fstride*in_stride;
+        }while( (Fout += m) != Fout_end );
+    }
+
+    Fout=Fout_beg;
+
+    // recombine the p smaller DFTs 
+    switch (p) {
+        case 2: kf_bfly2(Fout,fstride,st,m); break;
+        case 3: kf_bfly3(Fout,fstride,st,m); break; 
+        case 4: kf_bfly4(Fout,fstride,st,m); break;
+        case 5: kf_bfly5(Fout,fstride,st,m); break; 
+        default: kf_bfly_generic(Fout,fstride,st,m,p); break;
+    }
+}
+
+/*  facbuf is populated by p1,m1,p2,m2, ...
+    where 
+    p[i] * m[i] = m[i-1]
+    m0 = n                  */
+static 
+void kf_factor(int n,int * facbuf)
+{
+    int p=4;
+    double floor_sqrt;
+    floor_sqrt = floor( sqrt((double)n) );
+
+    /*factor out powers of 4, powers of 2, then any remaining primes */
+    do {
+        while (n % p) {
+            switch (p) {
+                case 4: p = 2; break;
+                case 2: p = 3; break;
+                default: p += 2; break;
+            }
+            if (p > floor_sqrt)
+                p = n;          /* no more factors, skip to end */
+        }
+        n /= p;
+        *facbuf++ = p;
+        *facbuf++ = n;
+    } while (n > 1);
+}
+
+/*
+ *
+ * User-callable function to allocate all necessary storage space for the fft.
+ *
+ * The return value is a contiguous block of memory, allocated with malloc.  As such,
+ * It can be freed with free(), rather than a kiss_fft-specific function.
+ * */
+kiss_fft_cfg kiss_fft_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem )
+{
+    kiss_fft_cfg st=NULL;
+    size_t memneeded = sizeof(struct kiss_fft_state)
+        + sizeof(kiss_fft_cpx)*(nfft-1); /* twiddle factors*/
+
+    if ( lenmem==NULL ) {
+        st = ( kiss_fft_cfg)KISS_FFT_MALLOC( memneeded );
+    }else{
+        if (mem != NULL && *lenmem >= memneeded)
+            st = (kiss_fft_cfg)mem;
+        *lenmem = memneeded;
+    }
+    if (st) {
+        int i;
+        st->nfft=nfft;
+        st->inverse = inverse_fft;
+
+        for (i=0;i<nfft;++i) {
+            const double pi=3.141592653589793238462643383279502884197169399375105820974944;
+            double phase = -2*pi*i / nfft;
+            if (st->inverse)
+                phase *= -1;
+            kf_cexp(st->twiddles+i, phase );
+        }
+
+        kf_factor(nfft,st->factors);
+    }
+    return st;
+}
+
+
+void kiss_fft_stride(kiss_fft_cfg st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout,int in_stride)
+{
+    if (fin == fout) {
+        //NOTE: this is not really an in-place FFT algorithm.
+        //It just performs an out-of-place FFT into a temp buffer
+        kiss_fft_cpx * tmpbuf = (kiss_fft_cpx*)KISS_FFT_TMP_ALLOC( sizeof(kiss_fft_cpx)*st->nfft);
+        kf_work(tmpbuf,fin,1,in_stride, st->factors,st);
+        memcpy(fout,tmpbuf,sizeof(kiss_fft_cpx)*st->nfft);
+        KISS_FFT_TMP_FREE(tmpbuf);
+    }else{
+        kf_work( fout, fin, 1,in_stride, st->factors,st );
+    }
+}
+
+void kiss_fft(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout)
+{
+    kiss_fft_stride(cfg,fin,fout,1);
+}
+
+
+void kiss_fft_cleanup(void)
+{
+    // nothing needed any more
+}
+
+int kiss_fft_next_fast_size(int n)
+{
+    while(1) {
+        int m=n;
+        while ( (m%2) == 0 ) m/=2;
+        while ( (m%3) == 0 ) m/=3;
+        while ( (m%5) == 0 ) m/=5;
+        if (m<=1)
+            break; /* n is completely factorable by twos, threes, and fives */
+        n++;
+    }
+    return n;
+}

kiss_fft.h

@@ -0,0 +1,124 @@
+#ifndef KISS_FFT_H
+#define KISS_FFT_H
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <math.h>
+#include <string.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ ATTENTION!
+ If you would like a :
+ -- a utility that will handle the caching of fft objects
+ -- real-only (no imaginary time component ) FFT
+ -- a multi-dimensional FFT
+ -- a command-line utility to perform ffts
+ -- a command-line utility to perform fast-convolution filtering
+
+ Then see kfc.h kiss_fftr.h kiss_fftnd.h fftutil.c kiss_fastfir.c
+  in the tools/ directory.
+*/
+
+#ifdef USE_SIMD
+# include <xmmintrin.h>
+# define kiss_fft_scalar __m128
+#define KISS_FFT_MALLOC(nbytes) _mm_malloc(nbytes,16)
+#define KISS_FFT_FREE _mm_free
+#else	
+#define KISS_FFT_MALLOC malloc
+#define KISS_FFT_FREE free
+#endif	
+
+
+#ifdef FIXED_POINT
+#include <sys/types.h>	
+# if (FIXED_POINT == 32)
+#  define kiss_fft_scalar int32_t
+# else	
+#  define kiss_fft_scalar int16_t
+# endif
+#else
+# ifndef kiss_fft_scalar
+/*  default is float */
+#   define kiss_fft_scalar float
+# endif
+#endif
+
+typedef struct {
+    kiss_fft_scalar r;
+    kiss_fft_scalar i;
+}kiss_fft_cpx;
+
+typedef struct kiss_fft_state* kiss_fft_cfg;
+
+/* 
+ *  kiss_fft_alloc
+ *  
+ *  Initialize a FFT (or IFFT) algorithm's cfg/state buffer.
+ *
+ *  typical usage:      kiss_fft_cfg mycfg=kiss_fft_alloc(1024,0,NULL,NULL);
+ *
+ *  The return value from fft_alloc is a cfg buffer used internally
+ *  by the fft routine or NULL.
+ *
+ *  If lenmem is NULL, then kiss_fft_alloc will allocate a cfg buffer using malloc.
+ *  The returned value should be free()d when done to avoid memory leaks.
+ *  
+ *  The state can be placed in a user supplied buffer 'mem':
+ *  If lenmem is not NULL and mem is not NULL and *lenmem is large enough,
+ *      then the function places the cfg in mem and the size used in *lenmem
+ *      and returns mem.
+ *  
+ *  If lenmem is not NULL and ( mem is NULL or *lenmem is not large enough),
+ *      then the function returns NULL and places the minimum cfg 
+ *      buffer size in *lenmem.
+ * */
+
+kiss_fft_cfg kiss_fft_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem); 
+
+/*
+ * kiss_fft(cfg,in_out_buf)
+ *
+ * Perform an FFT on a complex input buffer.
+ * for a forward FFT,
+ * fin should be  f[0] , f[1] , ... ,f[nfft-1]
+ * fout will be   F[0] , F[1] , ... ,F[nfft-1]
+ * Note that each element is complex and can be accessed like
+    f[k].r and f[k].i
+ * */
+void kiss_fft(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout);
+
+/*
+ A more generic version of the above function. It reads its input from every Nth sample.
+ * */
+void kiss_fft_stride(kiss_fft_cfg cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout,int fin_stride);
+
+/* If kiss_fft_alloc allocated a buffer, it is one contiguous 
+   buffer and can be simply free()d when no longer needed*/
+#define kiss_fft_free free
+
+/*
+ Cleans up some memory that gets managed internally. Not necessary to call, but it might clean up 
+ your compiler output to call this before you exit.
+*/
+void kiss_fft_cleanup(void);
+	
+
+/*
+ * Returns the smallest integer k, such that k>=n and k has only "fast" factors (2,3,5)
+ */
+int kiss_fft_next_fast_size(int n);
+
+/* for real ffts, we need an even size */
+#define kiss_fftr_next_fast_size_real(n) \
+        (kiss_fft_next_fast_size( ((n)+1)>>1)<<1)
+
+#ifdef __cplusplus
+} 
+#endif
+
+#endif

kiss_fftr.c

@@ -0,0 +1,159 @@
+/*
+Copyright (c) 2003-2004, Mark Borgerding
+
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+    * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+    * Neither the author nor the names of any contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#include "kiss_fftr.h"
+#include "_kiss_fft_guts.h"
+
+struct kiss_fftr_state{
+    kiss_fft_cfg substate;
+    kiss_fft_cpx * tmpbuf;
+    kiss_fft_cpx * super_twiddles;
+#ifdef USE_SIMD
+    void * pad;
+#endif
+};
+
+kiss_fftr_cfg kiss_fftr_alloc(int nfft,int inverse_fft,void * mem,size_t * lenmem)
+{
+    int i;
+    kiss_fftr_cfg st = NULL;
+    size_t subsize, memneeded;
+
+    if (nfft & 1) {
+        fprintf(stderr,"Real FFT optimization must be even.\n");
+        return NULL;
+    }
+    nfft >>= 1;
+
+    kiss_fft_alloc (nfft, inverse_fft, NULL, &subsize);
+    memneeded = sizeof(struct kiss_fftr_state) + subsize + sizeof(kiss_fft_cpx) * ( nfft * 3 / 2);
+
+    if (lenmem == NULL) {
+        st = (kiss_fftr_cfg) KISS_FFT_MALLOC (memneeded);
+    } else {
+        if (*lenmem >= memneeded)
+            st = (kiss_fftr_cfg) mem;
+        *lenmem = memneeded;
+    }
+    if (!st)
+        return NULL;
+
+    st->substate = (kiss_fft_cfg) (st + 1); /*just beyond kiss_fftr_state struct */
+    st->tmpbuf = (kiss_fft_cpx *) (((char *) st->substate) + subsize);
+    st->super_twiddles = st->tmpbuf + nfft;
+    kiss_fft_alloc(nfft, inverse_fft, st->substate, &subsize);
+
+    for (i = 0; i < nfft/2; ++i) {
+        double phase =
+            -3.14159265358979323846264338327 * ((double) (i+1) / nfft + .5);
+        if (inverse_fft)
+            phase *= -1;
+        kf_cexp (st->super_twiddles+i,phase);
+    }
+    return st;
+}
+
+void kiss_fftr(kiss_fftr_cfg st,const kiss_fft_scalar *timedata,kiss_fft_cpx *freqdata)
+{
+    /* input buffer timedata is stored row-wise */
+    int k,ncfft;
+    kiss_fft_cpx fpnk,fpk,f1k,f2k,tw,tdc;
+
+    if ( st->substate->inverse) {
+        fprintf(stderr,"kiss fft usage error: improper alloc\n");
+        exit(1);
+    }
+
+    ncfft = st->substate->nfft;
+
+    /*perform the parallel fft of two real signals packed in real,imag*/
+    kiss_fft( st->substate , (const kiss_fft_cpx*)timedata, st->tmpbuf );
+    /* The real part of the DC element of the frequency spectrum in st->tmpbuf
+     * contains the sum of the even-numbered elements of the input time sequence
+     * The imag part is the sum of the odd-numbered elements
+     *
+     * The sum of tdc.r and tdc.i is the sum of the input time sequence. 
+     *      yielding DC of input time sequence
+     * The difference of tdc.r - tdc.i is the sum of the input (dot product) [1,-1,1,-1... 
+     *      yielding Nyquist bin of input time sequence
+     */
+ 
+    tdc.r = st->tmpbuf[0].r;
+    tdc.i = st->tmpbuf[0].i;
+    C_FIXDIV(tdc,2);
+    CHECK_OVERFLOW_OP(tdc.r ,+, tdc.i);
+    CHECK_OVERFLOW_OP(tdc.r ,-, tdc.i);
+    freqdata[0].r = tdc.r + tdc.i;
+    freqdata[ncfft].r = tdc.r - tdc.i;
+#ifdef USE_SIMD    
+    freqdata[ncfft].i = freqdata[0].i = _mm_set1_ps(0);
+#else
+    freqdata[ncfft].i = freqdata[0].i = 0;
+#endif
+
+    for ( k=1;k <= ncfft/2 ; ++k ) {
+        fpk    = st->tmpbuf[k]; 
+        fpnk.r =   st->tmpbuf[ncfft-k].r;
+        fpnk.i = - st->tmpbuf[ncfft-k].i;
+        C_FIXDIV(fpk,2);
+        C_FIXDIV(fpnk,2);
+
+        C_ADD( f1k, fpk , fpnk );
+        C_SUB( f2k, fpk , fpnk );
+        C_MUL( tw , f2k , st->super_twiddles[k-1]);
+
+        freqdata[k].r = HALF_OF(f1k.r + tw.r);
+        freqdata[k].i = HALF_OF(f1k.i + tw.i);
+        freqdata[ncfft-k].r = HALF_OF(f1k.r - tw.r);
+        freqdata[ncfft-k].i = HALF_OF(tw.i - f1k.i);
+    }
+}
+
+void kiss_fftri(kiss_fftr_cfg st,const kiss_fft_cpx *freqdata,kiss_fft_scalar *timedata)
+{
+    /* input buffer timedata is stored row-wise */
+    int k, ncfft;
+
+    if (st->substate->inverse == 0) {
+        fprintf (stderr, "kiss fft usage error: improper alloc\n");
+        exit (1);
+    }
+
+    ncfft = st->substate->nfft;
+
+    st->tmpbuf[0].r = freqdata[0].r + freqdata[ncfft].r;
+    st->tmpbuf[0].i = freqdata[0].r - freqdata[ncfft].r;
+    C_FIXDIV(st->tmpbuf[0],2);
+
+    for (k = 1; k <= ncfft / 2; ++k) {
+        kiss_fft_cpx fk, fnkc, fek, fok, tmp;
+        fk = freqdata[k];
+        fnkc.r = freqdata[ncfft - k].r;
+        fnkc.i = -freqdata[ncfft - k].i;
+        C_FIXDIV( fk , 2 );
+        C_FIXDIV( fnkc , 2 );
+
+        C_ADD (fek, fk, fnkc);
+        C_SUB (tmp, fk, fnkc);
+        C_MUL (fok, tmp, st->super_twiddles[k-1]);
+        C_ADD (st->tmpbuf[k],     fek, fok);
+        C_SUB (st->tmpbuf[ncfft - k], fek, fok);
+#ifdef USE_SIMD        
+        st->tmpbuf[ncfft - k].i *= _mm_set1_ps(-1.0);
+#else
+        st->tmpbuf[ncfft - k].i *= -1;
+#endif
+    }
+    kiss_fft (st->substate, st->tmpbuf, (kiss_fft_cpx *) timedata);
+}

kiss_fftr.h

@@ -0,0 +1,46 @@
+#ifndef KISS_FTR_H
+#define KISS_FTR_H
+
+#include "kiss_fft.h"
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+    
+/* 
+ 
+ Real optimized version can save about 45% cpu time vs. complex fft of a real seq.
+
+ 
+ 
+ */
+
+typedef struct kiss_fftr_state *kiss_fftr_cfg;
+
+
+kiss_fftr_cfg kiss_fftr_alloc(int nfft,int inverse_fft,void * mem, size_t * lenmem);
+/*
+ nfft must be even
+
+ If you don't care to allocate space, use mem = lenmem = NULL 
+*/
+
+
+void kiss_fftr(kiss_fftr_cfg cfg,const kiss_fft_scalar *timedata,kiss_fft_cpx *freqdata);
+/*
+ input timedata has nfft scalar points
+ output freqdata has nfft/2+1 complex points
+*/
+
+void kiss_fftri(kiss_fftr_cfg cfg,const kiss_fft_cpx *freqdata,kiss_fft_scalar *timedata);
+/*
+ input freqdata has  nfft/2+1 complex points
+ output timedata has nfft scalar points
+*/
+
+#define kiss_fftr_free free
+
+#ifdef __cplusplus
+}
+#endif
+#endif

pico_sdk_import.cmake

@@ -0,0 +1,62 @@
+# This is a copy of <PICO_SDK_PATH>/external/pico_sdk_import.cmake
+
+# This can be dropped into an external project to help locate this SDK
+# It should be include()ed prior to project()
+
+if (DEFINED ENV{PICO_SDK_PATH} AND (NOT PICO_SDK_PATH))
+    set(PICO_SDK_PATH $ENV{PICO_SDK_PATH})
+    message("Using PICO_SDK_PATH from environment ('${PICO_SDK_PATH}')")
+endif ()
+
+if (DEFINED ENV{PICO_SDK_FETCH_FROM_GIT} AND (NOT PICO_SDK_FETCH_FROM_GIT))
+    set(PICO_SDK_FETCH_FROM_GIT $ENV{PICO_SDK_FETCH_FROM_GIT})
+    message("Using PICO_SDK_FETCH_FROM_GIT from environment ('${PICO_SDK_FETCH_FROM_GIT}')")
+endif ()
+
+if (DEFINED ENV{PICO_SDK_FETCH_FROM_GIT_PATH} AND (NOT PICO_SDK_FETCH_FROM_GIT_PATH))
+    set(PICO_SDK_FETCH_FROM_GIT_PATH $ENV{PICO_SDK_FETCH_FROM_GIT_PATH})
+    message("Using PICO_SDK_FETCH_FROM_GIT_PATH from environment ('${PICO_SDK_FETCH_FROM_GIT_PATH}')")
+endif ()
+
+set(PICO_SDK_PATH "${PICO_SDK_PATH}" CACHE PATH "Path to the PICO SDK")
+set(PICO_SDK_FETCH_FROM_GIT "${PICO_SDK_FETCH_FROM_GIT}" CACHE BOOL "Set to ON to fetch copy of PICO SDK from git if not otherwise locatable")
+set(PICO_SDK_FETCH_FROM_GIT_PATH "${PICO_SDK_FETCH_FROM_GIT_PATH}" CACHE FILEPATH "location to download SDK")
+
+if (NOT PICO_SDK_PATH)
+    if (PICO_SDK_FETCH_FROM_GIT)
+        include(FetchContent)
+        set(FETCHCONTENT_BASE_DIR_SAVE ${FETCHCONTENT_BASE_DIR})
+        if (PICO_SDK_FETCH_FROM_GIT_PATH)
+            get_filename_component(FETCHCONTENT_BASE_DIR "${PICO_SDK_FETCH_FROM_GIT_PATH}" REALPATH BASE_DIR "${CMAKE_SOURCE_DIR}")
+        endif ()
+        FetchContent_Declare(
+                pico_sdk
+                GIT_REPOSITORY https://github.com/raspberrypi/pico-sdk
+                GIT_TAG master
+        )
+        if (NOT pico_sdk)
+            message("Downloading PICO SDK")
+            FetchContent_Populate(pico_sdk)
+            set(PICO_SDK_PATH ${pico_sdk_SOURCE_DIR})
+        endif ()
+        set(FETCHCONTENT_BASE_DIR ${FETCHCONTENT_BASE_DIR_SAVE})
+    else ()
+        message(FATAL_ERROR
+                "PICO SDK location was not specified. Please set PICO_SDK_PATH or set PICO_SDK_FETCH_FROM_GIT to on to fetch from git."
+                )
+    endif ()
+endif ()
+
+get_filename_component(PICO_SDK_PATH "${PICO_SDK_PATH}" REALPATH BASE_DIR "${CMAKE_BINARY_DIR}")
+if (NOT EXISTS ${PICO_SDK_PATH})
+    message(FATAL_ERROR "Directory '${PICO_SDK_PATH}' not found")
+endif ()
+
+set(PICO_SDK_INIT_CMAKE_FILE ${PICO_SDK_PATH}/pico_sdk_init.cmake)
+if (NOT EXISTS ${PICO_SDK_INIT_CMAKE_FILE})
+    message(FATAL_ERROR "Directory '${PICO_SDK_PATH}' does not appear to contain the PICO SDK")
+endif ()
+
+set(PICO_SDK_PATH ${PICO_SDK_PATH} CACHE PATH "Path to the PICO SDK" FORCE)
+
+include(${PICO_SDK_INIT_CMAKE_FILE})

ws2812.c

@@ -0,0 +1,108 @@
+/**
+ * Copyright (c) 2020 Raspberry Pi (Trading) Ltd.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "pico/stdlib.h"
+#include "hardware/pio.h"
+#include "hardware/clocks.h"
+#include "ws2812.pio.h"
+
+#define IS_RGBW true
+#define NUM_PIXELS 150
+
+#ifdef PICO_DEFAULT_WS2812_PIN
+#define WS2812_PIN PICO_DEFAULT_WS2812_PIN
+#else
+// default to pin 2 if the board doesn't have a default WS2812 pin defined
+#define WS2812_PIN 2
+#endif
+
+static inline void put_pixel(uint32_t pixel_grb) {
+    pio_sm_put_blocking(pio0, 0, pixel_grb << 8u);
+}
+
+static inline uint32_t urgb_u32(uint8_t r, uint8_t g, uint8_t b) {
+return
+((uint32_t) (r) << 8) |
+((uint32_t) (g) << 16) |
+(uint32_t) (b);
+}
+
+void pattern_snakes(uint len, uint t) {
+    for (uint i = 0; i < len; ++i) {
+        uint x = (i + (t >> 1)) % 64;
+        if (x < 10)
+            put_pixel(urgb_u32(0xff, 0, 0));
+        else if (x >= 15 && x < 25)
+            put_pixel(urgb_u32(0, 0xff, 0));
+        else if (x >= 30 && x < 40)
+            put_pixel(urgb_u32(0, 0, 0xff));
+        else
+            put_pixel(0);
+    }
+}
+
+void pattern_random(uint len, uint t) {
+    if (t % 8)
+        return;
+    for (int i = 0; i < len; ++i)
+        put_pixel(rand());
+}
+
+void pattern_sparkle(uint len, uint t) {
+    if (t % 8)
+        return;
+    for (int i = 0; i < len; ++i)
+        put_pixel(rand() % 16 ? 0 : 0xffffffff);
+}
+
+void pattern_greys(uint len, uint t) {
+    int max = 100; // let's not draw too much current!
+    t %= max;
+    for (int i = 0; i < len; ++i) {
+        put_pixel(t * 0x10101);
+        if (++t >= max) t = 0;
+    }
+}
+
+typedef void (*pattern)(uint len, uint t);
+const struct {
+    pattern pat;
+    const char *name;
+} pattern_table[] = {
+        {pattern_snakes,  "Snakes!"},
+        {pattern_random,  "Random data"},
+        {pattern_sparkle, "Sparkles"},
+        {pattern_greys,   "Greys"},
+};
+
+//int main() {
+//    //set_sys_clock_48();
+//    stdio_init_all();
+//    printf("WS2812 Smoke Test, using pin %d", WS2812_PIN);
+//
+//    // todo get free sm
+//    PIO pio = pio0;
+//    int sm = 0;
+//    uint offset = pio_add_program(pio, &ws2812_program);
+//
+//    ws2812_program_init(pio, sm, offset, WS2812_PIN, 800000, IS_RGBW);
+//
+//    int t = 0;
+//    while (1) {
+//        int pat = rand() % count_of(pattern_table);
+//        int dir = (rand() >> 30) & 1 ? 1 : -1;
+//        puts(pattern_table[pat].name);
+//        puts(dir == 1 ? "(forward)" : "(backward)");
+//        for (int i = 0; i < 1000; ++i) {
+//            pattern_table[pat].pat(NUM_PIXELS, t);
+//            sleep_ms(10);
+//            t += dir;
+//        }
+//    }
+//}

ws2812.pio

@@ -0,0 +1,85 @@
+;
+; Copyright (c) 2020 Raspberry Pi (Trading) Ltd.
+;
+; SPDX-License-Identifier: BSD-3-Clause
+;
+
+.program ws2812
+.side_set 1
+
+.define public T1 2
+.define public T2 5
+.define public T3 3
+
+.lang_opt python sideset_init = pico.PIO.OUT_HIGH
+.lang_opt python out_init     = pico.PIO.OUT_HIGH
+.lang_opt python out_shiftdir = 1
+
+.wrap_target
+bitloop:
+    out x, 1       side 0 [T3 - 1] ; Side-set still takes place when instruction stalls
+    jmp !x do_zero side 1 [T1 - 1] ; Branch on the bit we shifted out. Positive pulse
+do_one:
+    jmp  bitloop   side 1 [T2 - 1] ; Continue driving high, for a long pulse
+do_zero:
+    nop            side 0 [T2 - 1] ; Or drive low, for a short pulse
+.wrap
+
+% c-sdk {
+#include "hardware/clocks.h"
+
+static inline void ws2812_program_init(PIO pio, uint sm, uint offset, uint pin, float freq, bool rgbw) {
+
+    pio_gpio_init(pio, pin);
+    pio_sm_set_consecutive_pindirs(pio, sm, pin, 1, true);
+
+    pio_sm_config c = ws2812_program_get_default_config(offset);
+    sm_config_set_sideset_pins(&c, pin);
+    sm_config_set_out_shift(&c, false, true, rgbw ? 32 : 24);
+    sm_config_set_fifo_join(&c, PIO_FIFO_JOIN_TX);
+
+    int cycles_per_bit = ws2812_T1 + ws2812_T2 + ws2812_T3;
+    float div = clock_get_hz(clk_sys) / (freq * cycles_per_bit);
+    sm_config_set_clkdiv(&c, div);
+
+    pio_sm_init(pio, sm, offset, &c);
+    pio_sm_set_enabled(pio, sm, true);
+}
+%}
+
+.program ws2812_parallel
+
+.define public T1 2
+.define public T2 5
+.define public T3 3
+
+.wrap_target
+    out x, 32
+    mov pins, !null [T1-1]
+    mov pins, x     [T2-1]
+    mov pins, null  [T3-2]
+.wrap
+
+% c-sdk {
+#include "hardware/clocks.h"
+
+static inline void ws2812_parallel_program_init(PIO pio, uint sm, uint offset, uint pin_base, uint pin_count, float freq) {
+    for(uint i=pin_base; i<pin_base+pin_count; i++) {
+        pio_gpio_init(pio, i);
+    }
+    pio_sm_set_consecutive_pindirs(pio, sm, pin_base, pin_count, true);
+
+    pio_sm_config c = ws2812_parallel_program_get_default_config(offset);
+    sm_config_set_out_shift(&c, true, true, 32);
+    sm_config_set_out_pins(&c, pin_base, pin_count);
+    sm_config_set_set_pins(&c, pin_base, pin_count);
+    sm_config_set_fifo_join(&c, PIO_FIFO_JOIN_TX);
+
+    int cycles_per_bit = ws2812_parallel_T1 + ws2812_parallel_T2 + ws2812_parallel_T3;
+    float div = clock_get_hz(clk_sys) / (freq * cycles_per_bit);
+    sm_config_set_clkdiv(&c, div);
+
+    pio_sm_init(pio, sm, offset, &c);
+    pio_sm_set_enabled(pio, sm, true);
+}
+%}