lstm_gpu_kernel.h 8.6 KB
Newer Older
D
dangqingqing 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.

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. */

#pragma once
16 17
#include <type_traits>
#include "paddle/operators/math/detail/hl_activation_functions.h"
D
dangqingqing 已提交
18 19
#include "paddle/operators/math/lstm_compute.h"
#include "paddle/platform/cuda_helper.h"
20
#include "paddle/platform/device_context.h"
D
dangqingqing 已提交
21

22 23
#include <glog/logging.h>

D
dangqingqing 已提交
24 25 26 27 28 29 30 31 32 33
namespace paddle {
namespace operators {
namespace math {
namespace detail {

/*
 * threads(framePerBlock, batchPerBlock)
 * grid(frameBlocks, batchBlocks)
 */
template <class T, class Op, bool isBatch>
34 35 36 37
__global__ void KeLstmForward(Op op, LstmMetaValue<T> value, int frameSize,
                              int batchSize, activation_mode_t active_node,
                              activation_mode_t active_gate,
                              activation_mode_t active_state) {
D
dangqingqing 已提交
38 39 40 41 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 67 68 69 70 71 72
  const int frameIdx = blockIdx.x * blockDim.x + threadIdx.x;
  if (frameIdx >= frameSize) return;

  int batchIdx = 0;
  if (isBatch) {
    batchIdx = blockIdx.y * blockDim.y + threadIdx.y;
    if (batchIdx >= batchSize) return;
    value.gateValue += batchIdx * frameSize * 4;
    value.outputValue += batchIdx * frameSize;
    value.stateValue += batchIdx * frameSize;
    value.stateActiveValue += batchIdx * frameSize;
  }

  T rState;
  T rPrevState = 0;
  T rStateAtv;
  T rOut;
  T rValueIn;
  T rValueIg;
  T rValueFg;
  T rValueOg;
  T rCheckI = value.checkIg[frameIdx];
  T rCheckF = value.checkFg[frameIdx];
  T rCheckO = value.checkOg[frameIdx];

  rValueIn = value.gateValue[frameIdx];
  rValueIg = value.gateValue[frameIdx + frameSize];
  rValueFg = value.gateValue[frameIdx + frameSize * 2];
  rValueOg = value.gateValue[frameIdx + frameSize * 3];

  if (value.prevStateValue) {
    if (isBatch) value.prevStateValue += batchIdx * frameSize;
    rPrevState = value.prevStateValue[frameIdx];
  }

73
  hppl::gpu::ForwardAct<T> act;
D
dangqingqing 已提交
74
  op(rValueIn, rValueIg, rValueFg, rValueOg, rPrevState, rState, rStateAtv,
75 76
     rOut, rCheckI, rCheckF, rCheckO, act(active_node), act(active_gate),
     act(active_state));
D
dangqingqing 已提交
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92

  value.gateValue[frameIdx] = rValueIn;
  value.gateValue[frameIdx + frameSize] = rValueIg;
  value.gateValue[frameIdx + frameSize * 2] = rValueFg;
  value.gateValue[frameIdx + frameSize * 3] = rValueOg;

  value.stateValue[frameIdx] = rState;
  value.stateActiveValue[frameIdx] = rStateAtv;
  value.outputValue[frameIdx] = rOut;
}

/*
 * threads(framePerBlock, batchPerBlock)
 * grid(frameBlocks, batchBlocks)
 */
template <class T, class Op, bool isBatch>
93 94 95 96 97
__global__ void KeLstmBackward(Op op, LstmMetaValue<T> value,
                               LstmMetaGrad<T> grad, int frameSize,
                               int batchSize, activation_mode_t active_node,
                               activation_mode_t active_gate,
                               activation_mode_t active_state) {
D
dangqingqing 已提交
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
  const int frameIdx = blockIdx.x * blockDim.x + threadIdx.x;
  if (frameIdx >= frameSize) return;

  int batchIdx = 0;
  if (isBatch) {
    batchIdx = blockIdx.y * blockDim.y + threadIdx.y;
    if (batchIdx >= batchSize) return;
    value.gateValue += batchIdx * frameSize * 4;
    value.stateValue += batchIdx * frameSize;
    value.stateActiveValue += batchIdx * frameSize;
    grad.gateGrad += batchIdx * frameSize * 4;
    grad.stateGrad += batchIdx * frameSize;
    grad.outputGrad += batchIdx * frameSize;
  }

  T rValueIn;
  T rValueIg;
  T rValueFg;
  T rValueOg;
  T rGradIn;
  T rGradIg;
  T rGradFg;
  T rGradOg;
  T rPrevState = 0;
  T rPrevStateGrad;
  T rState;
  T rStateGrad;
  T rStateAtv;
  T rOutputGrad;
  T rCheckI = value.checkIg[frameIdx];
  T rCheckF = value.checkFg[frameIdx];
  T rCheckO = value.checkOg[frameIdx];
  T rCheckIGrad;
  T rCheckFGrad;
  T rCheckOGrad;

  rValueIn = value.gateValue[frameIdx];
  rValueIg = value.gateValue[frameIdx + frameSize];
  rValueFg = value.gateValue[frameIdx + frameSize * 2];
  rValueOg = value.gateValue[frameIdx + frameSize * 3];
  rState = value.stateValue[frameIdx];
  rStateAtv = value.stateActiveValue[frameIdx];
  rOutputGrad = grad.outputGrad[frameIdx];
  rStateGrad = grad.stateGrad[frameIdx];

  if (value.prevStateValue) {
    if (isBatch) value.prevStateValue += batchIdx * frameSize;
    rPrevState = value.prevStateValue[frameIdx];
  }

148
  hppl::gpu::BackwardAct<T> act;
D
dangqingqing 已提交
149 150 151
  op(rValueIn, rValueIg, rValueFg, rValueOg, rGradIn, rGradIg, rGradFg, rGradOg,
     rPrevState, rPrevStateGrad, rState, rStateGrad, rStateAtv, rOutputGrad,
     rCheckI, rCheckF, rCheckO, rCheckIGrad, rCheckFGrad, rCheckOGrad,
152
     act(active_node), act(active_gate), act(active_state));
D
dangqingqing 已提交
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184

  grad.gateGrad[frameIdx] = rGradIn;
  grad.gateGrad[frameIdx + frameSize] = rGradIg;
  grad.gateGrad[frameIdx + frameSize * 2] = rGradFg;
  grad.gateGrad[frameIdx + frameSize * 3] = rGradOg;
  grad.stateGrad[frameIdx] = rStateGrad;
  if (grad.prevStateGrad) {
    if (isBatch) grad.prevStateGrad += batchIdx * frameSize;
    grad.prevStateGrad[frameIdx] = rPrevStateGrad;
  }

  if (isBatch) {
    if (value.prevStateValue) {
      if (grad.checkIgGrad)
        paddle::platform::CudaAtomicAdd(grad.checkIgGrad + frameIdx,
                                        rCheckIGrad);
      if (grad.checkFgGrad)
        paddle::platform::CudaAtomicAdd(grad.checkFgGrad + frameIdx,
                                        rCheckFGrad);
    }
    if (grad.checkOgGrad)
      paddle::platform::CudaAtomicAdd(grad.checkOgGrad + frameIdx, rCheckOGrad);
  } else {
    if (value.prevStateValue) {
      if (grad.checkIgGrad) grad.checkIgGrad[frameIdx] += rCheckIGrad;
      if (grad.checkFgGrad) grad.checkFgGrad[frameIdx] += rCheckFGrad;
    }
    if (grad.checkOgGrad) grad.checkOgGrad[frameIdx] += rCheckOGrad;
  }
}

template <class T, class Op>
185 186
void gpu_lstm_forward(const platform::DeviceContext& context, Op op,
                      LstmMetaValue<T> value, int frameSize, int batchSize,
D
dangqingqing 已提交
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202
                      activation_mode_t active_node,
                      activation_mode_t active_gate,
                      activation_mode_t active_state) {
  dim3 threads;
  dim3 grid;
  if (batchSize == 1) {
    int framePerBlock = frameSize <= 1024 ? frameSize : 1024;
    int frameBlocks = (frameSize + 1024 - 1) / 1024;
    threads = dim3(framePerBlock, 1);
    grid = dim3(frameBlocks, 1);
  } else {
    /* framePerBlock = 32 batchPerBlock = 32 */
    threads = dim3(32, 32);
    grid = dim3((frameSize + 32 - 1) / 32, (batchSize + 32 - 1) / 32);
  }

203 204
  auto stream =
      reinterpret_cast<const platform::CUDADeviceContext&>(context).stream();
D
dangqingqing 已提交
205 206
  if (batchSize == 1) {
    KeLstmForward<T, Op,
207
                  /* isBatch= */ false><<<grid, threads, 0, stream>>>(
208
        op, value, frameSize, batchSize, active_node, active_gate, active_gate);
D
dangqingqing 已提交
209 210
  } else {
    KeLstmForward<T, Op,
211
                  /* isBatch= */ true><<<grid, threads, 0, stream>>>(
212
        op, value, frameSize, batchSize, active_node, active_gate, active_gate);
D
dangqingqing 已提交
213 214 215 216
  }
}

template <class T, class Op>
217 218 219 220
void gpu_lstm_backward(const platform::DeviceContext& context, Op op,
                       LstmMetaValue<T> value, LstmMetaGrad<T> grad,
                       int frameSize, int batchSize,
                       activation_mode_t active_node,
D
dangqingqing 已提交
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
                       activation_mode_t active_gate,
                       activation_mode_t active_state) {
  dim3 threads;
  dim3 grid;
  if (batchSize == 1) {
    int framePerBlock = frameSize <= 1024 ? frameSize : 1024;
    int frameBlocks = (frameSize + 1024 - 1) / 1024;
    threads = dim3(framePerBlock, 1);
    grid = dim3(frameBlocks, 1);
  } else {
    /* framePerBlock = 32 batchPerBlock = 32 */
    threads = dim3(32, 32);
    grid = dim3((frameSize + 32 - 1) / 32, (batchSize + 32 - 1) / 32);
  }

236 237
  auto stream =
      reinterpret_cast<const platform::CUDADeviceContext&>(context).stream();
D
dangqingqing 已提交
238 239
  if (batchSize == 1) {
    KeLstmBackward<T, Op,
240
                   /* isBatch= */ false><<<grid, threads, 0, stream>>>(
241 242
        op, value, grad, frameSize, batchSize, active_node, active_gate,
        active_state);
D
dangqingqing 已提交
243 244
  } else {
    KeLstmBackward<T, Op,
245
                   /* isBatch= */ true><<<grid, threads, 0, stream>>>(
246 247
        op, value, grad, frameSize, batchSize, active_node, active_gate,
        active_state);
D
dangqingqing 已提交
248 249 250 251 252 253 254
  }
}

}  // namespace detail
}  // namespace math
}  // namespace operators
}  // namespace paddle