Update SIMD version of life demo.

native_client_sdk/src/examples/demo/life_simd/life.cc

+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <assert.h>
+#include <math.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/time.h>
+#include <unistd.h>
+
+#include <ppapi/c/ppb_input_event.h>
+#include <ppapi/cpp/fullscreen.h>
+#include <ppapi/cpp/input_event.h>
+#include <ppapi/cpp/var.h>
+#include <ppapi/cpp/var_array.h>
+#include <ppapi/cpp/var_array_buffer.h>
+#include <ppapi/cpp/var_dictionary.h>
+
+#include "ppapi_simple/ps.h"
+#include "ppapi_simple/ps_context_2d.h"
+#include "ppapi_simple/ps_event.h"
+#include "ppapi_simple/ps_instance.h"
+#include "ppapi_simple/ps_interface.h"
+#include "ppapi_simple/ps_main.h"
+#include "sdk_util/macros.h"
+#include "sdk_util/thread_pool.h"
+
+using namespace sdk_util;  // For sdk_util::ThreadPool
+
+namespace {
+
+#define INLINE inline __attribute__((always_inline))
+
+// BGRA helper macro, for constructing a pixel for a BGRA buffer.
+#define MakeBGRA(b, g, r, a)  \
+  (((a) << 24) | ((r) << 16) | ((g) << 8) | (b))
+
+const int kFramesToBenchmark = 100;
+const int kCellAlignment = 0x10;
+
+// 128 bit vector types
+typedef uint8_t u8x16_t __attribute__ ((vector_size (16)));
+
+// Helper function to broadcast x across 16 element vector.
+INLINE u8x16_t broadcast(uint8_t x) {
+  u8x16_t r = {x, x, x, x, x, x, x, x, x, x, x, x, x, x, x, x};
+  return r;
+}
+
+// Convert a count value into a live (green) or dead color value.
+const uint32_t kNeighborColors[] = {
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0xFF, 0x00, 0xFF),
+    MakeBGRA(0x00, 0xFF, 0x00, 0xFF),
+    MakeBGRA(0x00, 0xFF, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+    MakeBGRA(0x00, 0x00, 0x00, 0xFF),
+};
+
+// These represent the new health value of a cell based on its neighboring
+// values.  The health is binary: either alive or dead.
+const uint8_t kIsAlive[] = {
+      0, 0, 0, 0, 0, 1, 1, 1, 0,
+      0, 0, 0, 0, 0, 0, 0, 0, 0
+};
+
+// Timer helper for benchmarking.  Returns seconds elapsed since program start,
+// as a double.
+timeval start_tv;
+int start_tv_retv = gettimeofday(&start_tv, NULL);
+
+inline double getseconds() {
+  const double usec_to_sec = 0.000001;
+  timeval tv;
+  if ((0 == start_tv_retv) && (0 == gettimeofday(&tv, NULL)))
+    return (tv.tv_sec - start_tv.tv_sec) + tv.tv_usec * usec_to_sec;
+  return 0.0;
+}
+} // namespace
+
+
+class Life {
+ public:
+  Life();
+  virtual ~Life();
+  // Runs a tick of the simulations, update 2D output.
+  void Update();
+  // Handle event from user, or message from JS.
+  void HandleEvent(PSEvent* ps_event);
+ private:
+  void UpdateContext();
+  void DrawCell(int32_t x, int32_t y);
+  void ProcessTouchEvent(pp::TouchInputEvent touches);

[binji, 2014/08/08 00:31:40]

const pp::TouchInputEvent&

[nfullagar, 2014/08/08 20:51:48]

Done.

+  void PostUpdateMessage(const char* message, double value);
+  void StartBenchmark();
+  void EndBenchmark();
+  void Stir();
+  void wSimulate(int y);
+  static void wSimulateEntry(int y, void* data);
+  void Simulate();
+
+  bool simd_;
+  bool multithread_;
+  bool benchmarking_;
+  int benchmark_frame_counter_;
+  double bench_start_time_;
+  double bench_end_time_;
+  uint8_t* cell_in_;
+  uint8_t* cell_out_;
+  int32_t cell_stride_;
+  int32_t width_;
+  int32_t height_;
+  PSContext2D_t* ps_context_;
+  ThreadPool* workers_;
+};
+
+Life::Life() :
+    simd_(true),
+    multithread_(true),
+    benchmarking_(false),
+    benchmark_frame_counter_(0),
+    bench_start_time_(0.0),
+    bench_end_time_(0.0),
+    cell_in_(NULL),
+    cell_out_(NULL),
+    cell_stride_(0),
+    width_(0),
+    height_(0) {
+  ps_context_ = PSContext2DAllocate(PP_IMAGEDATAFORMAT_BGRA_PREMUL);
+  // Query system for number of processors via sysconf()
+  int num_threads = sysconf(_SC_NPROCESSORS_ONLN);
+  if (num_threads < 2)
+    num_threads = 2;
+  workers_ = new ThreadPool(num_threads);
+  PSEventSetFilter(PSE_ALL);
+}
+
+Life::~Life() {
+  delete workers_;
+  PSContext2DFree(ps_context_);
+}
+
+void Life::UpdateContext() {
+  cell_stride_ = (ps_context_->width + kCellAlignment - 1) &
+      ~(kCellAlignment - 1);

[binji, 2014/08/08 00:31:40]

whoops, bug in the previous version, hm?

+  size_t size = cell_stride_ * ps_context_->height;
+
+  if (ps_context_->width != width_ || ps_context_->height != height_) {
+    free(cell_in_);
+    free(cell_out_);
+
+    // Create a new context
+    void* in_buffer = NULL;
+    void* out_buffer = NULL;
+    // alloc buffers aligned on 16 bytes
+    posix_memalign(&in_buffer, kCellAlignment, size);
+    posix_memalign(&out_buffer, kCellAlignment, size);
+    cell_in_ = (uint8_t*) in_buffer;
+    cell_out_ = (uint8_t*) out_buffer;
+
+    memset(cell_out_, 0, size);
+    for (size_t index = 0; index < size; index++) {
+      cell_in_[index] = rand() & 1;
+    }
+    width_ = ps_context_->width;
+    height_ = ps_context_->height;
+  }
+}
+
+void Life::DrawCell(int32_t x, int32_t y) {
+  if (!cell_in_) return;
+  if (x > 0 && x < ps_context_->width - 1 &&
+      y > 0 && y < ps_context_->height - 1) {
+    cell_in_[x - 1 + y * cell_stride_] = 1;
+    cell_in_[x + 1 + y * cell_stride_] = 1;
+    cell_in_[x + (y - 1) * cell_stride_] = 1;
+    cell_in_[x + (y + 1) * cell_stride_] = 1;
+  }
+}
+
+void Life::ProcessTouchEvent(pp::TouchInputEvent touches) {
+  uint32_t count = touches.GetTouchCount(PP_TOUCHLIST_TYPE_TOUCHES);
+  uint32_t i, j;
+  for (i = 0; i < count; i++) {
+    pp::TouchPoint touch =
+        touches.GetTouchByIndex(PP_TOUCHLIST_TYPE_TOUCHES, i);
+    int radius = (int)(touch.radii().x());
+    int x = (int)(touch.position().x());
+    int y = (int)(touch.position().y());
+    // num = 1/100th the area of touch point
+    uint32_t num = (uint32_t)(M_PI * radius * radius / 100.0f);
+    for (j = 0; j < num; j++) {
+      int dx = rand() % (radius * 2) - radius;
+      int dy = rand() % (radius * 2) - radius;
+      // only plot random cells within the touch area
+      if (dx * dx + dy * dy <= radius * radius)
+        DrawCell(x + dx, y + dy);
+    }
+  }
+}
+
+void Life::PostUpdateMessage(const char* message_name, double value) {
+  pp::VarDictionary message;
+  message.Set("message", message_name);
+  message.Set("value", value);
+  PSInterfaceMessaging()->PostMessage(PSGetInstanceId(), message.pp_var());
+}
+
+void Life::StartBenchmark() {
+  printf("Running benchmark... (SIMD: %s, multi-threading: %s, size: %dx%d)\n",
+      simd_ ? "enabled" : "disabled",
+      multithread_ ? "enabled" : "disabled",
+      ps_context_->width,
+      ps_context_->height);
+  benchmarking_ = true;
+  bench_start_time_ = getseconds();
+  benchmark_frame_counter_ = kFramesToBenchmark;
+}
+
+void Life::EndBenchmark() {
+  double total_time;
+  bench_end_time_ = getseconds();
+  benchmarking_ = false;
+  total_time = bench_end_time_ - bench_start_time_;
+  printf("Finished - benchmark took %f seconds\n", total_time);
+  // Send benchmark result to JS.
+  PostUpdateMessage("benchmark_result", total_time);
+}
+
+void Life::HandleEvent(PSEvent* ps_event) {
+  // Give the 2D context a chance to process the event.
+  if (0 != PSContext2DHandleEvent(ps_context_, ps_event)) {
+    UpdateContext();
+    return;
+  }
+
+  switch(ps_event->type) {
+
+    case PSE_INSTANCE_HANDLEINPUT: {
+      pp::InputEvent event(ps_event->as_resource);
+
+      switch(event.GetType()) {
+        case PP_INPUTEVENT_TYPE_MOUSEDOWN:
+        case PP_INPUTEVENT_TYPE_MOUSEMOVE: {
+          pp::MouseInputEvent mouse = pp::MouseInputEvent(event);
+          // If the button is down, draw
+          if (mouse.GetModifiers() & PP_INPUTEVENT_MODIFIER_LEFTBUTTONDOWN) {
+            PP_Point location = mouse.GetPosition();
+            DrawCell(location.x, location.y);
+          }
+          break;
+        }
+
+        case PP_INPUTEVENT_TYPE_TOUCHSTART:
+        case PP_INPUTEVENT_TYPE_TOUCHMOVE: {
+          pp::TouchInputEvent touches = pp::TouchInputEvent(event);
+          ProcessTouchEvent(touches);
+          break;
+        }
+
+        case PP_INPUTEVENT_TYPE_KEYDOWN: {
+          pp::Fullscreen fullscreen(PSInstance::GetInstance());
+          bool isFullscreen = fullscreen.IsFullscreen();
+          fullscreen.SetFullscreen(!isFullscreen);
+          break;
+        }
+
+        default:
+          break;
+      }
+      break;  // case PSE_INSTANCE_HANDLEINPUT
+    }
+
+    case PSE_INSTANCE_HANDLEMESSAGE: {
+      // Convert Pepper Simple message to PPAPI C++ vars
+      pp::Var var(ps_event->as_var);
+      if (var.is_dictionary()) {
+        pp::VarDictionary dictionary(var);
+        std::string message = dictionary.Get("message").AsString();
+        if (message == "run_benchmark" && !benchmarking_) {
+          StartBenchmark();
+        } else if (message == "set_simd") {
+          std::string value = dictionary.Get("value").AsString();
+          simd_ = value == "enable";
+        } else if (message == "set_threading") {
+          std::string value = dictionary.Get("value").AsString();
+          multithread_ = value == "enable";
+        }
+      }
+      break;  // case PSE_INSTANCE_HANDLEMESSAGE
+    }
+
+    default:
+      break;
+  }
+}
+
+void Life::Stir() {
+  int32_t width = ps_context_->width;
+  int32_t height = ps_context_->height;
+  int32_t stride = cell_stride_;
+  int32_t i;
+  if (cell_in_ == NULL || cell_out_ == NULL)
+    return;
+
+  for (i = 0; i < width; ++i) {
+    cell_in_[i] = rand() & 1;
+    cell_in_[i + (height - 1) * stride] = rand() & 1;
+  }
+  for (i = 0; i < height; ++i) {
+    cell_in_[i * stride] = rand() & 1;
+    cell_in_[i * stride + (width - 1)] = rand() & 1;
+  }
+}
+
+void Life::wSimulate(int y) {
+  // Don't run simulation on top and bottom borders
+  if (y < 1 || y >= ps_context_->height - 1)
+    return;
+
+  // Do neighbor summation; apply rules, output pixel color. Note that a 1 cell
+  // wide perimeter is excluded from the simulation update; only cells from
+  // x = 1 to x < width - 1 and y = 1 to y < height - 1 are updated.
+  uint8_t *src0 = (cell_in_ + (y - 1) * cell_stride_);
+  uint8_t *src1 = src0 + cell_stride_;
+  uint8_t *src2 = src1 + cell_stride_;
+  uint8_t *dst = (cell_out_ + y * cell_stride_) + 1;
+  uint32_t *pixels = static_cast<uint32_t *>(ps_context_->data);
+  uint32_t *pixel_line = static_cast<uint32_t*>

[binji, 2014/08/08 00:31:40]

cast isn't needed, right?

+      (pixels + y * ps_context_->stride / sizeof(uint32_t));
+  int32_t x = 1;
+
+  if (simd_) {
+    const u8x16_t kOne = broadcast(1);
+    const u8x16_t kFour = broadcast(4);
+    const u8x16_t kEight = broadcast(8);
+    const u8x16_t kZero255 = {0, 255, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
+
+    // Prime the src
+    u8x16_t src00 = *reinterpret_cast<u8x16_t*>(&src0[0]);
+    u8x16_t src01 = *reinterpret_cast<u8x16_t*>(&src0[16]);
+    u8x16_t src10 = *reinterpret_cast<u8x16_t*>(&src1[0]);
+    u8x16_t src11 = *reinterpret_cast<u8x16_t*>(&src1[16]);
+    u8x16_t src20 = *reinterpret_cast<u8x16_t*>(&src2[0]);
+    u8x16_t src21 = *reinterpret_cast<u8x16_t*>(&src2[16]);
+
+    // This inner loop is SIMD - each loop iteration will process 16 cells.
+    for (; (x + 15) < (ps_context_->width - 1); x += 16) {
+
+      // Construct jittered source temps, using __builtin_shufflevector(..) to
+      // extract a shifted 16 element vector from the 32 element concatenation
+      // of two source vectors.
+      u8x16_t src0j0 = src00;
+      u8x16_t src0j1 = __builtin_shufflevector(src00, src01,
+          1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16);
+      u8x16_t src0j2 = __builtin_shufflevector(src00, src01,
+          2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17);
+      u8x16_t src1j0 = src10;
+      u8x16_t src1j1 = __builtin_shufflevector(src10, src11,
+          1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16);
+      u8x16_t src1j2 = __builtin_shufflevector(src10, src11,
+          2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17);
+      u8x16_t src2j0 = src20;
+      u8x16_t src2j1 = __builtin_shufflevector(src20, src21,
+          1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16);
+      u8x16_t src2j2 = __builtin_shufflevector(src20, src21,
+          2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17);
+
+      // Sum the jittered sources to construct neighbor count.
+      u8x16_t count = src0j0 + src0j1 +  src0j2 +
+                      src1j0 +        +  src1j2 +
+                      src2j0 + src2j1 +  src2j2;
+      // Add the center cell.
+      count = count + count + src1j1;
+      // If count > 4 and < 8, center cell will be alive in the next frame.
+      u8x16_t alive1 = count > kFour;
+      u8x16_t alive2 = count < kEight;
+      // Intersect the two comparisons from above.
+      u8x16_t alive = alive1 & alive2;
+
+      // At this point, alive[x] will be one of two values:
+      //   0x00 for a dead cell
+      //   0xFF for an alive cell.
+      //
+      // Next, convert alive cells to green pixel color.
+      // Use __builtin_shufflevector(..) to construct output pixels from
+      // concantination of alive vector and kZero255 const vector.
+      //   Indices 0..15 select the 16 cells from alive vector.
+      //   Index 16 is zero constant from kZero255 constant vector.
+      //   Index 17 is 255 constant from kZero255 constant vector.
+      //   Output pixel color values are in BGRABGRABGRABGRA order.
+      // Since each pixel needs 4 bytes of color information, 16 cells will
+      // need to expand to 4 seperate 16 byte pixel splats.
+      u8x16_t pixel0_3 = __builtin_shufflevector(alive, kZero255,
+        16, 0, 16, 17, 16, 1, 16, 17, 16, 2, 16, 17, 16, 3, 16, 17);
+      u8x16_t pixel4_7 = __builtin_shufflevector(alive, kZero255,
+        16, 4, 16, 17, 16, 5, 16, 17, 16, 6, 16, 17, 16, 7, 16, 17);
+      u8x16_t pixel8_11 = __builtin_shufflevector(alive, kZero255,
+        16, 8, 16, 17, 16, 9, 16, 17, 16, 10, 16, 17, 16, 11, 16, 17);
+      u8x16_t pixel12_15 = __builtin_shufflevector(alive, kZero255,
+        16, 12, 16, 17, 16, 13, 16, 17, 16, 14, 16, 17, 16, 15, 16, 17);
+
+      // Write 16 pixels to output pixel buffer.
+      *reinterpret_cast<u8x16_t*>(pixel_line + 0) = pixel0_3;
+      *reinterpret_cast<u8x16_t*>(pixel_line + 4) = pixel4_7;
+      *reinterpret_cast<u8x16_t*>(pixel_line + 8) = pixel8_11;
+      *reinterpret_cast<u8x16_t*>(pixel_line + 12) = pixel12_15;
+
+      // Convert alive mask to 1 or 0 and store in destination cell array.
+      *reinterpret_cast<u8x16_t*>(dst) = alive & kOne;
+
+      // Increment pointers.
+      pixel_line += 16;
+      dst += 16;
+      src0 += 16;
+      src1 += 16;
+      src2 += 16;
+
+      // Shift source over by 16 cells and read the next 16 cells.
+      src00 = src01;
+      src01 = *reinterpret_cast<u8x16_t*>(&src0[16]);
+      src10 = src11;
+      src11 = *reinterpret_cast<u8x16_t*>(&src1[16]);
+      src20 = src21;
+      src21 = *reinterpret_cast<u8x16_t*>(&src2[16]);
+    }
+  }
+
+  // The SIMD loop above does 16 cells at a time.  The loop below is the
+  // regular version which processes one cell at a time.  It is used to
+  // finish the remainder of the scanline not handled by the SIMD loop.
+  for (; x < (ps_context_->width - 1); ++x) {
+    // Sum the jittered sources to construct neighbor count.
+    int count = src0[0] + src0[1] + src0[2] +
+                src1[0] +         + src1[2] +
+                src2[0] + src2[1] + src2[2];
+    // Add the center cell.
+    count = count + count + src1[1];
+    // Use table lookup indexed by count to determine pixel & alive state.
+    uint32_t color = kNeighborColors[count];
+    *pixel_line++ = color;
+    *dst++ = kIsAlive[count];
+    ++src0;
+    ++src1;
+    ++src2;
+  }
+}
+
+// Static entry point for worker thread.
+void Life::wSimulateEntry(int slice, void* thiz) {
+  static_cast<Life*>(thiz)->wSimulate(slice);
+}
+
+void Life::Simulate() {
+  // Stir up the edges to prevent the simulation from reaching steady state.
+  Stir();
+
+  if (multithread_) {
+    // If multi-threading enabled, dispatch tasks to pool of worker threads.
+    workers_->Dispatch(ps_context_->height, wSimulateEntry, this);
+  } else {
+    // Else manually simulate each line on this thread.
+    for (int y = 0; y < ps_context_->height; y++) {
+      wSimulateEntry(y, this);
+    }
+  }
+  std::swap(cell_in_, cell_out_);
+}
+
+void Life::Update() {
+
+  PSContext2DGetBuffer(ps_context_);
+  if (NULL == ps_context_->data)
+    return;
+
+  // If we somehow have not allocated these pointers yet, skip this frame.
+  if (!cell_in_ || !cell_out_) return;
+
+  // Simulate one (or more if benchmarking) frames
+  do {
+    Simulate();
+    if (!benchmarking_)
+      break;
+    --benchmark_frame_counter_;
+  } while(benchmark_frame_counter_ > 0);
+  if (benchmarking_)
+    EndBenchmark();
+
+ PSContext2DSwapBuffer(ps_context_);
+}
+
+// Starting point for the module.  We do not use main since it would
+// collide with main in libppapi_cpp.
+int example_main(int argc, char* argv[]) {
+  Life life;
+  while (true) {
+    PSEvent* ps_event;
+    // Consume all available events
+    while ((ps_event = PSEventTryAcquire()) != NULL) {
+      life.HandleEvent(ps_event);
+      PSEventRelease(ps_event);
+    }
+    // Do simulation, render and present.
+    life.Update();
+  }
+  return 0;
+}
+
+// Register the function to call once the Instance Object is initialized.
+// see: pappi_simple/ps_main.h
+PPAPI_SIMPLE_REGISTER_MAIN(example_main);