blas_impl.cu.h 14.9 KB
Newer Older
Y
Yu Yang 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
//   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// 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

#include "paddle/fluid/operators/math/math_function.h"
#include "paddle/fluid/platform/dynload/cublas.h"
19 20 21
#include "paddle/fluid/platform/gpu_info.h"

DECLARE_bool(enable_cublas_tensor_op_math);
Y
Yu Yang 已提交
22 23 24 25 26 27 28 29 30 31 32 33 34 35

namespace paddle {
namespace operators {
namespace math {

template <typename T>
struct CUBlas;

template <>
struct CUBlas<float> {
  template <typename... ARGS>
  static void GEMM(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasSgemm(args...));
  }
Y
Yu Yang 已提交
36 37 38 39 40 41 42 43 44 45 46 47

  template <typename... ARGS>
  static void AXPY(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasSaxpy(args...));
  }

  template <typename... ARGS>
  static void GEMV(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasSgemv(args...));
  }

  template <typename... ARGS>
48
  static void GEMM_STRIDED_BATCH(ARGS... args) {
Y
Yu Yang 已提交
49 50 51 52
#if CUDA_VERSION >= 8000
    PADDLE_ENFORCE(platform::dynload::cublasSgemmStridedBatched(args...));
#else
    PADDLE_THROW("SgemmStridedBatched is not supported on cuda <= 7.5");
53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85
#endif
  }

  // NOTES: GEMM_EX can use Tensor Core to accelerate matrix multiply.
  // https://docs.nvidia.com/cuda/cublas/index.html#cublassetmathmode
  template <typename... ARGS>
  static void GEMM_EX(platform::CUDADeviceContext *dev_ctx,
                      cublasOperation_t transa, cublasOperation_t transb, int m,
                      int n, int k, const float *alpha, const void *A,
                      cudaDataType_t Atype, int lda, const void *B,
                      cudaDataType_t Btype, int ldb, const float *beta, void *C,
                      cudaDataType_t Ctype, int ldc) {
    // Because the gcc 4.8 doesn't expand template parameter pack that
    // appears in a lambda-expression, I can not use template parameter pack
    // here.
    auto cublas_call = [&]() {
#if CUDA_VERSION >= 8000
      VLOG(5) << "use_tensor_op_math: "
              << (platform::TensorCoreAvailable() ? "True" : "False");
      PADDLE_ENFORCE(platform::dynload::cublasSgemmEx(
          dev_ctx->cublas_handle(), transa, transb, m, n, k, alpha, A, Atype,
          lda, B, Btype, ldb, beta, C, Ctype, ldc));
#else
      PADDLE_THROW("cublasSgemmEx is supported on cuda >= 8.0");
#endif
    };

#if CUDA_VERSION >= 9000
    // NOTES: To use Tensor Core, we should change the cublas config,
    // but the cublas may be hold by multi-thread.
    dev_ctx->CublasCall(cublas_call, CUBLAS_TENSOR_OP_MATH);
#else
    cublas_call();
Y
Yu Yang 已提交
86 87
#endif
  }
Y
Yu Yang 已提交
88 89 90 91 92 93 94 95
};

template <>
struct CUBlas<double> {
  template <typename... ARGS>
  static void GEMM(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasDgemm(args...));
  }
Y
Yu Yang 已提交
96 97 98 99 100 101 102 103 104 105 106 107

  template <typename... ARGS>
  static void AXPY(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasDaxpy(args...));
  }

  template <typename... ARGS>
  static void GEMV(ARGS... args) {
    PADDLE_ENFORCE(platform::dynload::cublasDgemv(args...));
  }

  template <typename... ARGS>
108
  static void GEMM_STRIDED_BATCH(ARGS... args) {
Y
Yu Yang 已提交
109 110 111 112 113 114
#if CUDA_VERSION >= 8000
    PADDLE_ENFORCE(platform::dynload::cublasDgemmStridedBatched(args...));
#else
    PADDLE_THROW("DgemmStridedBatched is not supported on cuda <= 7.5");
#endif
  }
115 116 117 118 119

  template <typename... ARGS>
  static void GEMM_EX(ARGS... args) {
    PADDLE_THROW("Currently there are not cublasDgemmEx.");
  }
Y
Yu Yang 已提交
120 121 122 123
};

template <>
struct CUBlas<platform::float16> {
Y
Yu Yang 已提交
124 125 126 127 128 129 130 131 132 133 134 135 136 137
  using float16 = platform::float16;

  static void GEMM(cublasHandle_t handle, cublasOperation_t transa,
                   cublasOperation_t transb, int m, int n, int k,
                   const float16 *alpha, const float16 *A, int lda,
                   const float16 *B, int ldb, const float16 *beta, float16 *C,
                   int ldc) {
    PADDLE_ENFORCE(
        platform::dynload::cublasHgemm(handle, transa, transb, m, n, k,
                                       reinterpret_cast<const __half *>(alpha),
                                       reinterpret_cast<const __half *>(A), lda,
                                       reinterpret_cast<const __half *>(B), ldb,
                                       reinterpret_cast<const __half *>(beta),
                                       reinterpret_cast<__half *>(C), ldc));
Y
Yu Yang 已提交
138
  }
Y
Yu Yang 已提交
139

140 141 142 143 144 145 146 147 148 149
  static void GEMM_STRIDED_BATCH(cublasHandle_t handle,
                                 cublasOperation_t transa,
                                 cublasOperation_t transb, int m, int n, int k,
                                 const float16 *alpha, const float16 *A,
                                 int lda, long long int strideA,  // NOLINT
                                 const float16 *B,                // NOLINT
                                 int ldb, long long int strideB,  // NOLINT
                                 const float16 *beta, float16 *C, int ldc,
                                 long long int strideC,  // NOLINT
                                 int batchCount) {
Y
Yu Yang 已提交
150
#if CUDA_VERSION >= 8000
Y
yuyang18 已提交
151 152 153 154 155 156 157
    PADDLE_ENFORCE(platform::dynload::cublasHgemmStridedBatched(
        handle, transa, transb, m, n, k,
        reinterpret_cast<const __half *>(alpha),
        reinterpret_cast<const __half *>(A), lda, strideA,
        reinterpret_cast<const __half *>(B), ldb, strideB,
        reinterpret_cast<const __half *>(beta), reinterpret_cast<__half *>(C),
        ldc, strideC, batchCount));
Y
Yu Yang 已提交
158 159
#else
    PADDLE_THROW("HgemmStridedBatched is not supported on cuda <= 7.5");
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
#endif
  }

  // NOTES: GEMM_EX can use Tensor Core to accelerate matrix multiply.
  // https://docs.nvidia.com/cuda/cublas/index.html#cublassetmathmode
  template <typename... ARGS>
  static void GEMM_EX(platform::CUDADeviceContext *dev_ctx,
                      cublasOperation_t transa, cublasOperation_t transb, int m,
                      int n, int k, const void *alpha, const void *A,
                      cudaDataType_t Atype, int lda, const void *B,
                      cudaDataType_t Btype, int ldb, const void *beta, void *C,
                      cudaDataType_t Ctype, int ldc,
                      cudaDataType_t computeType) {
    auto cublas_call = [&]() {
#if CUDA_VERSION >= 8000
      cublasGemmAlgo_t algo = CUBLAS_GEMM_DFALT;
#if CUDA_VERSION >= 9000
      bool use_tensor_op_math = platform::TensorCoreAvailable();
      if (use_tensor_op_math) {
        algo = CUBLAS_GEMM_DFALT_TENSOR_OP;
      }
      VLOG(5) << "use_tensor_op_math: "
              << (use_tensor_op_math ? "True" : "False");
#endif  // CUDA_VERSION >= 9000

      PADDLE_ENFORCE(platform::dynload::cublasGemmEx(
          dev_ctx->cublas_handle(), transa, transb, m, n, k, alpha, A, Atype,
          lda, B, Btype, ldb, beta, C, Ctype, ldc, computeType, algo));
#else
      PADDLE_THROW("cublasGemmEx is supported on cuda >= 8.0");
#endif
    };

#if CUDA_VERSION >= 9000
    // NOTES: To use Tensor Core, we should change the cublas config,
    // but the cublas may be hold by multi-thread.
    dev_ctx->CublasCall(cublas_call, CUBLAS_TENSOR_OP_MATH);
#else
    cublas_call();
Y
Yu Yang 已提交
199 200
#endif
  }
Y
Yu Yang 已提交
201 202 203 204
};

template <>
template <typename T>
Y
Yu Yang 已提交
205 206 207 208
void Blas<platform::CUDADeviceContext>::GEMM(CBLAS_TRANSPOSE transA,
                                             CBLAS_TRANSPOSE transB, int M,
                                             int N, int K, T alpha, const T *A,
                                             const T *B, T beta, T *C) const {
Y
Yu Yang 已提交
209 210 211 212 213 214 215 216 217
  // Note that cublas follows fortran order, so the order is different from
  // the cblas convention.
  int lda = (transA == CblasNoTrans) ? K : M;
  int ldb = (transB == CblasNoTrans) ? N : K;
  cublasOperation_t cuTransA =
      (transA == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;
  cublasOperation_t cuTransB =
      (transB == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;

218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
#if CUDA_VERSION >= 8000
  if (FLAGS_enable_cublas_tensor_op_math && std::is_same<T, float>::value) {
    auto &cuda_ctx = const_cast<platform::CUDADeviceContext &>(context_);
    CUBlas<T>::GEMM_EX(&cuda_ctx, cuTransB, cuTransA, N, M, K, &alpha, B,
                       CUDA_R_32F, ldb, A, CUDA_R_32F, lda, &beta, C,
                       CUDA_R_32F, N);
  } else {
#endif  // CUDA_VERSION >= 8000

    CUBlas<T>::GEMM(context_.cublas_handle(), cuTransB, cuTransA, N, M, K,
                    &alpha, B, ldb, A, lda, &beta, C, N);

#if CUDA_VERSION >= 8000
  }
#endif  // CUDA_VERSION >= 8000
Y
Yu Yang 已提交
233 234 235 236 237
}

template <>
template <>
inline void Blas<platform::CUDADeviceContext>::GEMM(
Y
Yu Yang 已提交
238 239 240 241
    CBLAS_TRANSPOSE transA, CBLAS_TRANSPOSE transB, int M, int N, int K,
    platform::float16 alpha, const platform::float16 *A,
    const platform::float16 *B, platform::float16 beta,
    platform::float16 *C) const {
Y
Yu Yang 已提交
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
  // Note that cublas follows fortran order, so the order is different from
  // the cblas convention.
  int lda = (transA == CblasNoTrans) ? K : M;
  int ldb = (transB == CblasNoTrans) ? N : K;
  cublasOperation_t cuTransA =
      (transA == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;
  cublasOperation_t cuTransB =
      (transB == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;

  // TODO(kexinzhao): add processing code for compute capability < 53 case
  PADDLE_ENFORCE_GE(context_.GetComputeCapability(), 53,
                    "cublas fp16 gemm requires GPU compute capability >= 53");

  float h_alpha = static_cast<float>(alpha);
  float h_beta = static_cast<float>(beta);

258
#if CUDA_VERSION >= 8000
Y
Yu Yang 已提交
259 260 261 262
  // cublasHgemm does true FP16 computation which is slow for non-Volta
  // GPUs. So use cublasGemmEx instead which does pesudo FP16 computation:
  // input/output in fp16, computation in fp32, which can also be accelerated
  // using tensor cores in volta GPUs.
263 264 265 266
  auto &cuda_ctx = const_cast<platform::CUDADeviceContext &>(context_);
  CUBlas<platform::float16>::GEMM_EX(
      &cuda_ctx, cuTransB, cuTransA, N, M, K, &h_alpha, B, CUDA_R_16F, ldb, A,
      CUDA_R_16F, lda, &h_beta, C, CUDA_R_16F, N, CUDA_R_32F);
Y
Yu Yang 已提交
267 268
#else
  // CUDA 7.5 does not support cublasGemmEx, hence we fall back to use hgemm
Y
Yu Yang 已提交
269 270 271
  CUBlas<platform::float16>::GEMM(context_.cublas_handle(), cuTransB, cuTransA,
                                  N, M, K, &h_alpha, h_B, ldb, h_A, lda,
                                  &h_beta, h_C, N);
Y
Yu Yang 已提交
272 273 274 275 276
#endif  // CUDA_VERSION >= 8000
}

template <>
template <typename T>
Y
Yu Yang 已提交
277 278 279 280
void Blas<platform::CUDADeviceContext>::GEMM(bool transA, bool transB, int M,
                                             int N, int K, T alpha, const T *A,
                                             int lda, const T *B, int ldb,
                                             T beta, T *C, int ldc) const {
Y
Yu Yang 已提交
281 282
  // Note that cublas follows fortran order, so the order is different from
  // the cblas convention.
Y
Yu Yang 已提交
283 284
  cublasOperation_t cuTransA = transA ? CUBLAS_OP_T : CUBLAS_OP_N;
  cublasOperation_t cuTransB = transB ? CUBLAS_OP_T : CUBLAS_OP_N;
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316

#if CUDA_VERSION >= 8000
  if (FLAGS_enable_cublas_tensor_op_math && std::is_same<T, float>::value) {
    auto &cuda_ctx = const_cast<platform::CUDADeviceContext &>(context_);
    CUBlas<T>::GEMM_EX(&cuda_ctx, cuTransB, cuTransA, N, M, K, &alpha, B,
                       CUDA_R_32F, ldb, A, CUDA_R_32F, lda, &beta, C,
                       CUDA_R_32F, ldc);
  } else {
#endif  // CUDA_VERSION >= 8000

    CUBlas<T>::GEMM(context_.cublas_handle(), cuTransB, cuTransA, N, M, K,
                    &alpha, B, ldb, A, lda, &beta, C, ldc);

#if CUDA_VERSION >= 8000
  }
#endif  // CUDA_VERSION >= 8000
}

template <>
template <>
inline void Blas<platform::CUDADeviceContext>::GEMM(
    bool transA, bool transB, int M, int N, int K, platform::float16 alpha,
    const platform::float16 *A, int lda, const platform::float16 *B, int ldb,
    platform::float16 beta, platform::float16 *C, int ldc) const {
  // Note that cublas follows fortran order, so the order is different from
  // the cblas convention.
  cublasOperation_t cuTransA = transA ? CUBLAS_OP_T : CUBLAS_OP_N;
  cublasOperation_t cuTransB = transB ? CUBLAS_OP_T : CUBLAS_OP_N;

  CUBlas<platform::float16>::GEMM(context_.cublas_handle(), cuTransB, cuTransA,
                                  N, M, K, &alpha, B, ldb, A, lda, &beta, C,
                                  ldc);
Y
Yu Yang 已提交
317 318
}

Y
Yu Yang 已提交
319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
template <>
template <typename T>
void Blas<platform::CUDADeviceContext>::AXPY(int n, T alpha, const T *x,
                                             T *y) const {
  CUBlas<T>::AXPY(context_.cublas_handle(), n, &alpha, x, 1, y, 1);
}

template <>
template <typename T>
void Blas<platform::CUDADeviceContext>::GEMV(bool trans_a, int M, int N,
                                             T alpha, const T *A, const T *B,
                                             T beta, T *C) const {
  cublasOperation_t cuTransA = !trans_a ? CUBLAS_OP_T : CUBLAS_OP_N;

  CUBlas<T>::GEMV(context_.cublas_handle(), cuTransA, N, M, &alpha, A, N, B, 1,
                  &beta, C, 1);
}

template <>
template <typename T>
void Blas<platform::CUDADeviceContext>::BatchedGEMM(
    CBLAS_TRANSPOSE transA, CBLAS_TRANSPOSE transB, int M, int N, int K,
    T alpha, const T *A, const T *B, T beta, T *C, int batchCount,
    int64_t strideA, int64_t strideB) const {
  // Note that cublas follows fortran order, so the order is different from
  // the cblas convention.
  int lda = (transA == CblasNoTrans) ? K : M;
  int ldb = (transB == CblasNoTrans) ? N : K;
  int ldc = N;
  cublasOperation_t cuTransA =
      (transA == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;
  cublasOperation_t cuTransB =
      (transB == CblasNoTrans) ? CUBLAS_OP_N : CUBLAS_OP_T;
  const int64_t strideC = M * N;

354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
#if CUDA_VERSION >= 9010
  if (FLAGS_enable_cublas_tensor_op_math && std::is_same<T, float>::value) {
    auto cublas_call = [&]() {
      cublasGemmAlgo_t algo = CUBLAS_GEMM_DFALT;
      bool use_tensor_op_math = platform::TensorCoreAvailable();
      if (use_tensor_op_math) {
        algo = CUBLAS_GEMM_DFALT_TENSOR_OP;
      }
      VLOG(5) << "use_tensor_op_math: "
              << (use_tensor_op_math ? "True" : "False");

      PADDLE_ENFORCE(platform::dynload::cublasGemmStridedBatchedEx(
          context_.cublas_handle(), cuTransB, cuTransA, N, M, K, &alpha, B,
          CUDA_R_32F, ldb, strideB, A, CUDA_R_32F, lda, strideA, &beta, C,
          CUDA_R_32F, ldc, strideC, batchCount, CUDA_R_32F, algo));
    };
    auto &dev_ctx = const_cast<platform::CUDADeviceContext &>(context_);
    dev_ctx.CublasCall(cublas_call, CUBLAS_TENSOR_OP_MATH);
  } else {
#endif  // CUDA_VERSION >= 9010

    CUBlas<T>::GEMM_STRIDED_BATCH(context_.cublas_handle(), cuTransB, cuTransA,
                                  N, M, K, &alpha, B, ldb, strideB, A, lda,
                                  strideA, &beta, C, ldc, strideC, batchCount);

#if CUDA_VERSION >= 9010
  }
#endif  // CUDA_VERSION >= 9010
Y
Yu Yang 已提交
382 383
}

Y
Yu Yang 已提交
384 385 386
}  // namespace math
}  // namespace operators
}  // namespace paddle