gpu_context.cc 29.5 KB
Newer Older
W
Wilber 已提交
1
/* Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
2
Copyright (c) 2022 NVIDIA Corporation. All rights reserved.
W
Wilber 已提交
3 4 5 6 7 8 9 10 11 12 13 14

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. */
15

16
#include "paddle/phi/backends/gpu/gpu_context.h"
W
Wilber 已提交
17
#include <algorithm>
W
Wilber 已提交
18 19 20 21 22 23
#include <array>
#include <functional>
#include <future>
#include <memory>
#include <mutex>

24
#include "paddle/phi/api/ext/exception.h"
W
Wilber 已提交
25

26 27 28 29 30
#include "paddle/phi/backends/gpu/gpu_decls.h"
#include "paddle/phi/backends/gpu/gpu_info.h"
#include "paddle/phi/common/float16.h"
#include "paddle/phi/common/place.h"
#include "paddle/phi/core/allocator.h"
W
Wilber 已提交
31 32

#ifdef PADDLE_WITH_CUDA
33 34 35 36
#include "paddle/phi/backends/dynload/cublas.h"
#include "paddle/phi/backends/dynload/cudnn.h"
#include "paddle/phi/backends/dynload/cusolver.h"
#include "paddle/phi/backends/dynload/cusparse.h"
W
Wilber 已提交
37
#if !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
38
#include "paddle/phi/backends/dynload/nccl.h"
W
Wilber 已提交
39 40 41 42
#endif  // !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
#endif  // PADDLE_WITH_CUDA

#ifdef PADDLE_WITH_HIP
43 44
#include "paddle/phi/backends/dynload/miopen.h"
#include "paddle/phi/backends/dynload/rocblas.h"
W
Wilber 已提交
45
#if !defined(__APPLE__) && defined(PADDLE_WITH_RCCL)
46
#include "paddle/phi/backends/dynload/rccl.h"
W
Wilber 已提交
47 48 49 50 51 52 53
#endif  // !defined(__APPLE__) && defined(PADDLE_WITH_RCCL)
#endif  // PADDLE_WITH_HIP

// NOTE: The paddle framework should add WITH_EIGEN option to support compile
// without eigen.
#include "unsupported/Eigen/CXX11/Tensor"

54
// TODO(phi): remove fluid header.
W
Wilber 已提交
55 56
#include "paddle/fluid/platform/enforce.h"

57
namespace phi {
W
Wilber 已提交
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 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158

namespace internal {

class EigenGpuStreamDevice : public Eigen::StreamInterface {
 public:
  EigenGpuStreamDevice() : scratch_(nullptr), semaphore_(nullptr) {
    Eigen::initializeDeviceProp();
  }
  ~EigenGpuStreamDevice() override {}

  void Reinitialize(gpuStream_t cuda_stream,
                    Allocator* allocator,
                    GPUPlace place) {
    stream_ = cuda_stream;
    place_ = place;
    allocator_ = allocator;
    device_prop_ = &Eigen::m_deviceProperties[place.device];
  }

  const gpuStream_t& stream() const override { return stream_; }

  const gpuDeviceProp& deviceProperties() const override {
    return *device_prop_;
  }

  void* allocate(size_t num_bytes) const override {
    if (UNLIKELY(num_bytes == 0)) {
      return nullptr;
    }
    auto buf = allocator_->Allocate(num_bytes);
    VLOG(4) << "Eigen allocated at " << buf->ptr() << " requested "
            << num_bytes;
    void* retv = buf->ptr();
    {
      std::lock_guard<std::mutex> lock(mtx_);
      allocations_.emplace(retv, std::move(buf));
    }
    return retv;
  }

  void deallocate(void* buffer) const override {
    if (LIKELY(buffer)) {
      std::lock_guard<std::mutex> lock(mtx_);
      allocations_.erase(buffer);
    }
  }

  void* scratchpad() const override {
    if (scratch_ == NULL) {
      scratch_ = allocate(Eigen::kGpuScratchSize + sizeof(unsigned int));
    }
    return scratch_;
  }

  unsigned int* semaphore() const override {
    if (semaphore_ == NULL) {
      char* scratch = static_cast<char*>(scratchpad()) + Eigen::kGpuScratchSize;
      semaphore_ = reinterpret_cast<unsigned int*>(scratch);
#ifdef PADDLE_WITH_HIP
      PADDLE_ENFORCE_GPU_SUCCESS(
          hipMemsetAsync(semaphore_, 0, sizeof(unsigned int), stream_));
#else
      PADDLE_ENFORCE_GPU_SUCCESS(
          cudaMemsetAsync(semaphore_, 0, sizeof(unsigned int), stream_));
#endif
    }
    return semaphore_;
  }

 private:
  GPUPlace place_;
  gpuStream_t stream_;                // not owned;
  Allocator* allocator_;              // not owned;
  const gpuDeviceProp* device_prop_;  // not owned;
  mutable void* scratch_;
  mutable unsigned int* semaphore_;
  mutable std::mutex mtx_;  // to protect allocations_
  mutable std::unordered_map<void*, Allocator::AllocationPtr> allocations_;
};

#ifdef PADDLE_WITH_HIP
static void StreamCallbackFunc(gpuStream_t stream,
                               gpuError_t status,
                               void* user_data)
#endif
#ifdef PADDLE_WITH_CUDA
#if CUDA_VERSION >= 10000
    static void CUDART_CB StreamCallbackFunc(void* user_data)
#else
    static void CUDART_CB
    StreamCallbackFunc(cudaStream_t stream, cudaError_t status, void* user_data)
#endif
#endif
{
  std::unique_ptr<std::function<void()>> func(
      reinterpret_cast<std::function<void()>*>(user_data));
  (*func)();
}

}  // namespace internal

159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
void DnnWorkspaceHandle::RunFuncSync(
    const std::function<void(void*)>& cudnn_func,
    size_t required_workspace_bytes,
    bool use_cached_allocation) {
  bool need_realloc = required_workspace_bytes > WorkspaceSize();
  if (need_realloc && !use_cached_allocation) {
    void* workspace_ptr = nullptr;
    size_t size = ((required_workspace_bytes + 255) >> 8) << 8;
    std::lock_guard<std::mutex> guard(*mtx_);
#ifdef PADDLE_WITH_HIP
    auto status = hipMalloc(&workspace_ptr, size);
#else
    auto status = cudaMalloc(&workspace_ptr, size);
#endif
    if (status == gpuSuccess) {
      cudnn_func(workspace_ptr);
      phi::backends::gpu::GpuStreamSync(stream_);
#ifdef PADDLE_WITH_HIP
      PADDLE_ENFORCE_GPU_SUCCESS(hipFree(workspace_ptr));
#else
      PADDLE_ENFORCE_GPU_SUCCESS(cudaFree(workspace_ptr));
#endif
      return;
    }
  }

  RunFunc(cudnn_func, required_workspace_bytes);
  if (need_realloc) {
    // Release the workspace allocated in this running.
    ResetWorkspace();
  }
}

W
Wilber 已提交
192
void DnnWorkspaceHandle::ResetWorkspace() { allocation_ = nullptr; }
W
Wilber 已提交
193

W
Wilber 已提交
194 195 196 197 198 199
void DnnWorkspaceHandle::ReallocWorkspace(size_t required_workspace_bytes) {
  if (required_workspace_bytes <= WorkspaceSize()) return;
  // reset allocation first before re-allocate to save memory
  allocation_.reset();
  allocation_ = allocator_->Allocate(required_workspace_bytes);
}
W
Wilber 已提交
200 201 202 203 204 205 206 207 208

struct GPUContext::Impl {
  void Init() {
    owned_ = true;
    backends::gpu::GPUDeviceGuard guard(place_.device);
    InitGpuProperties();
    InitStream();
    InitEigenDevice();
    InitBlasHandle();
209
    InitBlasLtHandle();
W
Wilber 已提交
210 211 212 213 214 215 216 217 218 219 220 221
    InitDNNHandle();
    InitSolverHandle();
    InitSparseHandle();
    InitDnnWorkspace();
  }

  void PartialInitWithoutAllocator() {
    owned_ = true;
    backends::gpu::GPUDeviceGuard guard(place_.device);
    InitGpuProperties();
    InitStream();
    InitBlasHandle();
222
    InitBlasLtHandle();
W
Wilber 已提交
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
    InitDNNHandle();
    InitSolverHandle();
    InitSparseHandle();
  }

  void PartialInitWithAllocator() {
    owned_ = true;
    backends::gpu::GPUDeviceGuard guard(place_.device);
    InitEigenDevice();
    InitDnnWorkspace();
  }

  Impl() : place_(GPUPlace()) {}

  explicit Impl(const GPUPlace& place) : place_(place) {}

  ~Impl() {
    backends::gpu::GPUDeviceGuard guard(place_.device);
    DestoryInternalWorkspace();
    DestoryInternalEigenDevice();
    DestroyInternalSparseHandle();
    DestroyInternalSolverHandle();
    DestroyInternalDnnHandle();
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
    if (nccl_comm_) {
      PADDLE_ENFORCE_GPU_SUCCESS(dynload::ncclCommDestroy(nccl_comm_));
    }
#endif
    DestroyInternalBlasHandle();
252
    DestroyInternalBlasLtHandle();
W
Wilber 已提交
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
    DestoryInternalStream();
  }

  const Place& GetPlace() const { return place_; }

  bool IsTensorCoreAvailable() const {
    return blas_tensor_core_handle_ != nullptr;
  }

  void InitGpuProperties() {
    backends::gpu::GPUDeviceGuard guard(place_.GetDeviceId());
    compute_capability_ =
        backends::gpu::GetGPUComputeCapability(place_.GetDeviceId());
    multi_process_ = backends::gpu::GetGPUMultiProcessors(place_.GetDeviceId());
    max_threads_per_mp_ =
        backends::gpu::GetGPUMaxThreadsPerMultiProcessor(place_.GetDeviceId());
    max_grid_dim_size_ =
        backends::gpu::GetGpuMaxGridDimSize(place_.GetDeviceId());
    max_threads_per_block_ =
        backends::gpu::GetGPUMaxThreadsPerBlock(place_.GetDeviceId());
    driver_version_ = backends::gpu::GetGPUDriverVersion(place_.GetDeviceId());
    runtime_version_ =
        backends::gpu::GetGPURuntimeVersion(place_.GetDeviceId());

    // TODO(wilber): glog may be replaced in the future?
    LOG_FIRST_N(WARNING, 1)
        << "Please NOTE: device: " << static_cast<int>(place_.device)
        << ", GPU Compute Capability: " << compute_capability_ / 10 << "."
        << compute_capability_ % 10
        << ", Driver API Version: " << driver_version_ / 1000 << "."
        << (driver_version_ % 100) / 10
        << ", Runtime API Version: " << runtime_version_ / 1000 << "."
        << (runtime_version_ % 100) / 10;
#ifdef PADDLE_WITH_HIP
    size_t miopen_major, miopen_minor, miopen_patch;
    PADDLE_ENFORCE_GPU_SUCCESS(
        dynload::miopenGetVersion(&miopen_major, &miopen_minor, &miopen_patch));
    auto cudnn_dso_ver =
        (miopen_major * 1000 + miopen_minor * 10 + miopen_patch) / 10;
    auto compile_miopen_version = MIOPEN_VERSION / 10;
    if (cudnn_dso_ver < static_cast<size_t>(compile_miopen_version)) {
      LOG_FIRST_N(WARNING, 1)
          << "WARNING: device: " << static_cast<int>(place_.device)
          << ". The installed Paddle is compiled with MIOPEN "
          << compile_miopen_version / 100 << "." << compile_miopen_version % 100
          << ", but MIOPEN version in your machine is " << cudnn_dso_ver / 100
          << "." << cudnn_dso_ver % 100
          << ", which may cause serious incompatible bug. "
          << "Please recompile or reinstall Paddle with compatible MIOPEN "
             "version.";
    }
#else
    size_t cudnn_dso_ver = dynload::cudnnGetVersion();
    LOG_FIRST_N(WARNING, 1) << "device: " << static_cast<int>(place_.device)
                            << ", cuDNN Version: " << cudnn_dso_ver / 1000
                            << "." << (cudnn_dso_ver % 1000) / 100 << ".";

    // Check CUDA/CUDNN version compatiblity
    auto local_cuda_version =
        (driver_version_ / 1000) * 10 + (driver_version_ % 100) / 10;
    auto compile_cuda_version =
        (CUDA_VERSION / 1000) * 10 + (CUDA_VERSION % 100) / 10;
    if (local_cuda_version < compile_cuda_version) {
      LOG_FIRST_N(WARNING, 1)
          << "WARNING: device: " << static_cast<int>(place_.device)
          << ". The installed Paddle is compiled with CUDA "
          << compile_cuda_version / 10 << "." << compile_cuda_version % 10
          << ", but CUDA runtime version in your machine is "
          << local_cuda_version / 10 << "." << local_cuda_version % 10
          << ", which may cause serious incompatible bug. "
          << "Please recompile or reinstall Paddle with compatible CUDA "
             "version.";
    }
#endif
  }

  void InitDnnWorkspace() {
    PD_CHECK(allocator_ != nullptr,
             "the device allocator for gpu context is nullptr.");
332
    workspace_ = new DnnWorkspaceHandle(allocator_, stream_);
W
Wilber 已提交
333 334 335 336 337
  }

  void DestoryInternalWorkspace() {
    if (owned_ && workspace_ != nullptr) {
      delete workspace_;
338
      workspace_ = nullptr;
W
Wilber 已提交
339 340 341
    }
  }

W
Wilber 已提交
342 343 344 345 346 347 348 349
  // TODO(wilber): The return type is a pointer, to be modified later.
  // DnnWorkspaceHandle* GetDnnWorkspace() {
  //   PD_CHECK(workspace_ != nullptr, "the gpu cudnn workspace is nullptr.");
  //   return workspace_;
  // }
  DnnWorkspaceHandle GetDnnWorkspace() {
    PD_CHECK(allocator_ != nullptr,
             "the device allocator for gpu context is nullptr.");
350
    return DnnWorkspaceHandle(allocator_, stream_);
W
Wilber 已提交
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
  }

  void InitStream() {
#ifdef PADDLE_WITH_HIP
    PADDLE_ENFORCE_GPU_SUCCESS(
        hipStreamCreateWithPriority(&stream_, hipStreamDefault, 0));
#else
    PADDLE_ENFORCE_GPU_SUCCESS(
        cudaStreamCreateWithPriority(&stream_, cudaStreamDefault, 0));
#endif
  }

  void DestoryInternalStream() {
    if (owned_ && stream_ != nullptr) {
#ifdef PADDLE_WITH_HIP
      PADDLE_ENFORCE_GPU_SUCCESS(hipStreamDestroy(stream_));
#else
      PADDLE_ENFORCE_GPU_SUCCESS(cudaStreamDestroy(stream_));
#endif
    }
    stream_ = nullptr;
  }

  void SetStream(gpuStream_t stream) { stream_ = stream; }

  gpuStream_t GetStream() const {
    PD_CHECK(stream_ != nullptr, "the gpu stream is nullptr.");
    return stream_;
  }

  void InitEigenDevice() {
    PD_CHECK(allocator_ != nullptr,
             "the allocator for eigen device is nullptr.");
    eigen_stream_.reset(new internal::EigenGpuStreamDevice());
    eigen_stream_->Reinitialize(stream_, allocator_, place_);
    eigen_device_ = new Eigen::GpuDevice(eigen_stream_.get());
  }

  void DestoryInternalEigenDevice() {
    if (owned_ && eigen_device_ != nullptr) {
      delete eigen_device_;
      eigen_device_ = nullptr;
    }
  }

  void SetEigenDevice(Eigen::GpuDevice* device) { eigen_device_ = device; }

  Eigen::GpuDevice* eigen_device() const {
    PD_CHECK(eigen_device_ != nullptr, "the gpu eigen_device is nullptr.");
    return eigen_device_;
  }

  void InitBlasHandle() {
#ifdef PADDLE_WITH_HIP
405 406
    phi::dynload::rocblas_create_handle(&blas_handle_);
    phi::dynload::rocblas_set_stream(blas_handle_, stream_);
W
Wilber 已提交
407
#else  // PADDLE_WITH_CUDA
408
    PADDLE_RETRY_CUDA_SUCCESS(phi::dynload::cublasCreate(&blas_handle_));
W
Wilber 已提交
409
    PADDLE_RETRY_CUDA_SUCCESS(
410
        phi::dynload::cublasSetStream(blas_handle_, stream_));
W
Wilber 已提交
411 412
#if CUDA_VERSION >= 9000
    PADDLE_RETRY_CUDA_SUCCESS(
413
        phi::dynload::cublasCreate(&blas_tensor_core_handle_));
W
Wilber 已提交
414
    PADDLE_RETRY_CUDA_SUCCESS(
415 416
        phi::dynload::cublasSetStream(blas_tensor_core_handle_, stream_));
    PADDLE_RETRY_CUDA_SUCCESS(phi::dynload::cublasSetMathMode(
W
Wilber 已提交
417 418 419
        blas_tensor_core_handle_, CUBLAS_TENSOR_OP_MATH));
#if CUDA_VERSION >= 11000
    PADDLE_RETRY_CUDA_SUCCESS(
420
        phi::dynload::cublasCreate(&blas_tf32_tensor_core_handle_));
W
Wilber 已提交
421
    PADDLE_RETRY_CUDA_SUCCESS(
422 423
        phi::dynload::cublasSetStream(blas_tf32_tensor_core_handle_, stream_));
    PADDLE_RETRY_CUDA_SUCCESS(phi::dynload::cublasSetMathMode(
W
Wilber 已提交
424 425 426 427 428 429 430 431 432
        blas_tf32_tensor_core_handle_, CUBLAS_TF32_TENSOR_OP_MATH));
#endif  // CUDA_VERSION >= 11000
#endif  // CUDA_VERSION >= 9000
#endif  // PADDLE_WITH_HIP
  }

  void DestroyInternalBlasHandle() {
#ifdef PADDLE_WITH_HIP
    if (owned_ && blas_handle_ != nullptr) {
433
      phi::dynload::rocblas_destroy_handle(blas_handle_);
W
Wilber 已提交
434 435 436 437
      blas_handle_ = nullptr;
    }
#else
    if (owned_ && blas_handle_ != nullptr) {
438
      phi::dynload::cublasDestroy(blas_handle_);
W
Wilber 已提交
439 440 441
      blas_handle_ = nullptr;
    }
    if (owned_ && blas_tensor_core_handle_ != nullptr) {
442
      phi::dynload::cublasDestroy(blas_tensor_core_handle_);
W
Wilber 已提交
443 444 445
      blas_tensor_core_handle_ = nullptr;
    }
    if (owned_ && blas_tf32_tensor_core_handle_ != nullptr) {
446
      phi::dynload::cublasDestroy(blas_tf32_tensor_core_handle_);
W
Wilber 已提交
447 448 449 450 451 452 453 454 455 456 457 458
      blas_tf32_tensor_core_handle_ = nullptr;
    }
#endif  // PADDLE_WITH_HIP
  }

  blasHandle_t GetBlasHandle() const {
    PD_CHECK(blas_handle_ != nullptr, "the gpu blas handle is nullptr.");
    return blas_handle_;
  }

  void SetBlasHandle(blasHandle_t blas) { blas_handle_ = blas; }

459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
  void InitBlasLtHandle() {
#if defined(PADDLE_WITH_CUDA) && CUDA_VERSION >= 11060
    phi::dynload::cublasLtCreate(&blaslt_handle_);
#endif
  }

  void DestroyInternalBlasLtHandle() {
#if defined(PADDLE_WITH_CUDA) && CUDA_VERSION >= 11060
    phi::dynload::cublasLtDestroy(blaslt_handle_);
#endif
  }

  void SetBlasLtHandle(blasLtHandle_t blaslt) { blaslt_handle_ = blaslt; }

  blasLtHandle_t GetBlasLtHandle() const {
    PD_CHECK(blaslt_handle_ != nullptr, "the gpu blasLt handle is nullptr.");
    return blaslt_handle_;
  }

W
Wilber 已提交
478
  void InitDNNHandle() {
479
    if (phi::dynload::HasCUDNN()) {
W
Wilber 已提交
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
#ifdef PADDLE_WITH_HIP
      size_t miopen_major, miopen_minor, miopen_patch;
      PADDLE_ENFORCE_GPU_SUCCESS(dynload::miopenGetVersion(
          &miopen_major, &miopen_minor, &miopen_patch));
      auto local_miopen_version =
          (miopen_major * 1000 + miopen_minor * 10 + miopen_patch) / 10;
      auto compile_miopen_version = MIOPEN_VERSION / 10;
      if (local_miopen_version < static_cast<size_t>(compile_miopen_version)) {
        LOG_FIRST_N(WARNING, 1)
            << "WARNING: device: " << place_.device
            << ". The installed Paddle is compiled with MIOPEN "
            << compile_miopen_version / 100 << "."
            << compile_miopen_version % 100
            << ", but MIOPEN version in your machine is "
            << local_miopen_version / 100 << "." << local_miopen_version % 100
            << ", which may cause serious incompatible bug. "
            << "Please recompile or reinstall Paddle with compatible MIOPEN "
               "version.";
      }
      PADDLE_ENFORCE_GPU_SUCCESS(dynload::miopenCreate(&dnn_handle_));
      PADDLE_ENFORCE_GPU_SUCCESS(
          dynload::miopenSetStream(dnn_handle_, stream_));
#else
503
      auto local_cudnn_version = phi::dynload::cudnnGetVersion() / 100;
W
Wilber 已提交
504 505 506 507 508 509 510 511 512 513 514 515
      auto compile_cudnn_version = CUDNN_VERSION / 100;
      if (local_cudnn_version < static_cast<size_t>(compile_cudnn_version)) {
        LOG_FIRST_N(WARNING, 1)
            << "WARNING: device: " << place_.device
            << ". The installed Paddle is compiled with CUDNN "
            << compile_cudnn_version / 10 << "." << compile_cudnn_version % 10
            << ", but CUDNN version in your machine is "
            << local_cudnn_version / 10 << "." << local_cudnn_version % 10
            << ", which may cause serious incompatible bug. "
            << "Please recompile or reinstall Paddle with compatible CUDNN "
               "version.";
      }
516
      PADDLE_RETRY_CUDA_SUCCESS(phi::dynload::cudnnCreate(&dnn_handle_));
W
Wilber 已提交
517
      PADDLE_RETRY_CUDA_SUCCESS(
518
          phi::dynload::cudnnSetStream(dnn_handle_, stream_));
W
Wilber 已提交
519 520 521 522 523 524 525 526 527 528 529 530 531 532
#endif
    } else {
      dnn_handle_ = nullptr;
    }
  }

  dnnHandle_t GetDnnHandle() {
    PD_CHECK(dnn_handle_ != nullptr, "the gpu dnn handle is nullptr.");
    return dnn_handle_;
  }

  void DestroyInternalDnnHandle() {
#ifdef PADDLE_WITH_HIP
    if (owned_ && dnn_handle_ != nullptr) {
533
      PADDLE_ENFORCE_GPU_SUCCESS(phi::dynload::miopenDestroy(dnn_handle_));
W
Wilber 已提交
534 535 536 537
      dnn_handle_ = nullptr;
    }
#else
    if (owned_ && dnn_handle_ != nullptr) {
538
      PADDLE_ENFORCE_GPU_SUCCESS(phi::dynload::cudnnDestroy(dnn_handle_));
W
Wilber 已提交
539 540 541 542 543 544 545 546 547
      dnn_handle_ = nullptr;
    }
#endif  // PADDLE_WITH_HIP
  }

  void SetDnnHandle(dnnHandle_t handle) { dnn_handle_ = handle; }

  void InitSolverHandle() {
#ifndef PADDLE_WITH_HIP
548
    PADDLE_RETRY_CUDA_SUCCESS(phi::dynload::cusolverDnCreate(&solver_handle_));
W
Wilber 已提交
549
    PADDLE_RETRY_CUDA_SUCCESS(
550
        phi::dynload::cusolverDnSetStream(solver_handle_, stream_));
W
Wilber 已提交
551 552 553 554 555 556 557
#endif
  }

  void DestroyInternalSolverHandle() {
#ifndef PADDLE_WITH_HIP
    if (owned_ && solver_handle_ != nullptr) {
      PADDLE_ENFORCE_GPU_SUCCESS(
558
          phi::dynload::cusolverDnDestroy(solver_handle_));
W
Wilber 已提交
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
      solver_handle_ = nullptr;
    }
#endif
  }

  solverHandle_t GetSolverHandle() const {
    PD_CHECK(solver_handle_ != nullptr, "the gpu solver handle is nullptr.");
    return solver_handle_;
  }

  void SetSolverHandle(solverHandle_t handle) { solver_handle_ = handle; }

  void InitSparseHandle() {
// ROCM is not yet supported
#if defined(PADDLE_WITH_CUDA)
// The generic APIs is supported from CUDA10.1
#if CUDA_VERSION >= 10010
    PADDLE_RETRY_CUDA_SUCCESS(dynload::cusparseCreate(&sparse_handle_));
    PADDLE_RETRY_CUDA_SUCCESS(
        dynload::cusparseSetStream(sparse_handle_, stream_));
#endif
#endif
  }

  void DestroyInternalSparseHandle() {
#ifdef PADDLE_WITH_CUDA
#if CUDA_VERSION >= 10010
    if (owned_ && sparse_handle_ != nullptr) {
      PADDLE_RETRY_CUDA_SUCCESS(dynload::cusparseDestroy(sparse_handle_));
      sparse_handle_ = nullptr;
    }
#endif
#endif
  }

  sparseHandle_t GetSparseHandle() const {
    PD_CHECK(sparse_handle_ != nullptr, "the gpu sparse handle is nullptr.");
    return sparse_handle_;
  }

  void SetSparseHandle(sparseHandle_t handle) { sparse_handle_ = handle; }

  void Wait() const {
#ifdef PADDLE_WITH_HIP
    hipError_t e_sync = hipSuccess;
#if !defined(_WIN32)
    e_sync = hipStreamSynchronize(stream_);
#else
    while (e_sync = hipStreamQuery(stream_)) {
      if (e_sync == hipErrorNotReady) continue;
      break;
    }
#endif  // !defined(_WIN32)
#else   // PADDLE_WITH_HIP
    cudaError_t e_sync = cudaSuccess;
#if !defined(_WIN32)
    e_sync = cudaStreamSynchronize(stream_);
#else
    while (e_sync = cudaStreamQuery(stream_)) {
      if (e_sync == cudaErrorNotReady) continue;
      break;
    }
#endif  // !defined(_WIN32)
#endif  // PADDLE_WITH_HIP

    PADDLE_ENFORCE_GPU_SUCCESS(e_sync);
  }

  void WaitEvent(gpuEvent_t ev) const {
#ifdef PADDLE_WITH_HIP
    PADDLE_ENFORCE_GPU_SUCCESS(hipStreamWaitEvent(stream_, ev, 0));
#else
    PADDLE_ENFORCE_GPU_SUCCESS(cudaStreamWaitEvent(stream_, ev, 0));
#endif
  }

  ncclComm_t GetNcclComm() const {
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
    // PD_CHECK(nccl_comm_ != nullptr, "the gpu nccl_comm is nullptr.");
    return nccl_comm_;
#endif
    return nullptr;
  }

  void SetNcclComm(ncclComm_t comm) {
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
    nccl_comm_ = comm;
#endif
  }

  inline void CublasCall(
      const std::function<void(blasHandle_t)>& callback) const {
    if (blas_tf32_tensor_core_handle_ != nullptr) {
      std::lock_guard<std::mutex> guard(blas_tf32_mtx_);
      callback(blas_tf32_tensor_core_handle_);
    } else {
      std::lock_guard<std::mutex> guard(blas_mtx_);
      callback(blas_handle_);
    }
  }

  inline void TensorCoreCublasCallIfAvailable(
      const std::function<void(blasHandle_t)>& callback) const {
    if (blas_tensor_core_handle_ != nullptr) {
      std::lock_guard<std::mutex> guard(blas_tensor_core_mtx_);
      callback(blas_tensor_core_handle_);
    } else {
      std::lock_guard<std::mutex> guard(blas_mtx_);
      callback(blas_handle_);
    }
  }

  inline void CusparseCall(
      const std::function<void(sparseHandle_t)>& callback) const {
    std::lock_guard<std::mutex> guard(sparse_mtx_);
    callback(sparse_handle_);
  }

  void RecordEvent(gpuEvent_t ev, const std::function<void()>& callback) const {
    callback();
    RecordEvent(ev);
  }

  void RecordEvent(gpuEvent_t ev) const {
#ifdef PADDLE_WITH_HIP
    PADDLE_ENFORCE_GPU_SUCCESS(hipEventRecord(ev, stream_));
#else
    PADDLE_ENFORCE_GPU_SUCCESS(cudaEventRecord(ev, stream_));
#endif
  }

  void AddStreamCallback(const std::function<void()>& callback) const {
L
Leo Chen 已提交
691 692 693 694 695
    // NOTE(zhiqiu): better use threadpool here, otherwise "std::async" may
    // launch too
    // many threads and result in thread oversubscription.
    auto* callback_func = new std::function<void()>(std::move(callback));
    auto* func = new std::function<void()>([this, callback_func] {
W
Wilber 已提交
696
      std::lock_guard<std::mutex> lock(stream_call_back_mtx_);
L
Leo Chen 已提交
697 698 699 700 701
      VLOG(4) << "Stream callback";
      last_future_ = std::async(std::launch::async, [callback_func]() {
        std::unique_ptr<std::function<void()>> releaser(callback_func);
        (*callback_func)();
      });
W
Wilber 已提交
702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720
    });

#ifdef PADDLE_WITH_HIP
    PADDLE_ENFORCE_GPU_SUCCESS(
        hipStreamAddCallback(stream_, internal::StreamCallbackFunc, func, 0));
#endif
#ifdef PADDLE_WITH_CUDA
#if CUDA_VERSION >= 10000
    PADDLE_ENFORCE_GPU_SUCCESS(
        cudaLaunchHostFunc(stream_, internal::StreamCallbackFunc, func));
#else
    PADDLE_ENFORCE_GPU_SUCCESS(
        cudaStreamAddCallback(stream_, internal::StreamCallbackFunc, func, 0));
#endif
#endif
  }

  void WaitStreamCallback() const {
#if defined(PADDLE_WITH_HIP) || defined(PADDLE_WITH_CUDA)
721
    phi::backends::gpu::GpuStreamSync(stream_);
W
Wilber 已提交
722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745
#endif
    {
      std::lock_guard<std::mutex> lock(stream_call_back_mtx_);
      if (last_future_.valid()) {
        last_future_.wait();
      }
    }
  }

  bool owned_{false};
  Place place_;
  int compute_capability_;
  int runtime_version_;
  int driver_version_;
  int multi_process_;
  int max_threads_per_mp_;
  int max_threads_per_block_;
  std::array<int, 3> max_grid_dim_size_;

  gpuStream_t stream_{nullptr};
  Eigen::GpuDevice* eigen_device_{nullptr};
  blasHandle_t blas_handle_{nullptr};
  blasHandle_t blas_tensor_core_handle_{nullptr};
  blasHandle_t blas_tf32_tensor_core_handle_{nullptr};
746
  blasLtHandle_t blaslt_handle_{nullptr};
W
Wilber 已提交
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
  dnnHandle_t dnn_handle_{nullptr};
  solverHandle_t solver_handle_{nullptr};
  sparseHandle_t sparse_handle_{nullptr};
  DnnWorkspaceHandle* workspace_{nullptr};

#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
  // NCCL communicator (single process version) for NCCL collective operations.
  // NCCL collective operations provides fast collectives over multiple GPUs
  // both within and across nodes.
  // But, this collectives is used for collectives over multiple GPUs within
  // nodes.

  // NOTE: Distributed communicator, distributed framework manages its
  // resources.
  ncclComm_t nccl_comm_{nullptr};
#endif

  mutable std::mutex blas_mtx_;
  mutable std::mutex blas_tensor_core_mtx_;
  mutable std::mutex blas_tf32_mtx_;
  mutable std::mutex sparse_mtx_;
  mutable std::mutex stream_call_back_mtx_;
  mutable std::future<void> last_future_;

  Allocator* allocator_{nullptr};  // external resource.
  // A internal resouce to initinalize eigen_device.
  std::unique_ptr<internal::EigenGpuStreamDevice> eigen_stream_{nullptr};
};

GPUContext::GPUContext() : DeviceContext(), impl_(std::make_unique<Impl>()) {}

W
Wilber 已提交
778 779 780 781
GPUContext::GPUContext(GPUContext&&) = default;

GPUContext& GPUContext::operator=(GPUContext&&) = default;

W
Wilber 已提交
782 783 784 785 786 787 788 789 790 791 792 793 794 795 796
GPUContext::GPUContext(const GPUPlace& place)
    : DeviceContext(), impl_(std::make_unique<Impl>(place)) {}

GPUContext::~GPUContext() = default;

const Place& GPUContext::GetPlace() const { return impl_->GetPlace(); }

gpuStream_t GPUContext::stream() const { return impl_->GetStream(); }

dnnHandle_t GPUContext::cudnn_handle() const { return impl_->GetDnnHandle(); }

blasHandle_t GPUContext::cublas_handle() const {
  return impl_->GetBlasHandle();
}

797 798 799 800
blasLtHandle_t GPUContext::cublaslt_handle() const {
  return impl_->GetBlasLtHandle();
}

W
Wilber 已提交
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838
solverHandle_t GPUContext::cusolver_dn_handle() const {
  return impl_->GetSolverHandle();
}

sparseHandle_t GPUContext::cusparse_handle() const {
  return impl_->GetSparseHandle();
}

void GPUContext::Wait() const { impl_->Wait(); }

void GPUContext::WaitEvent(gpuEvent_t ev) const { impl_->WaitEvent(ev); }

bool GPUContext::tensor_core_available() const {
  return impl_->IsTensorCoreAvailable();
}

int GPUContext::GetComputeCapability() const {
  return impl_->compute_capability_;
}

int GPUContext::GetMaxPhysicalThreadCount() const {
  return impl_->multi_process_ * impl_->max_threads_per_mp_;
}

int GPUContext::GetSMCount() const { return impl_->multi_process_; }

int GPUContext::GetMaxThreadsPerBlock() const {
  return impl_->max_threads_per_block_;
}

std::array<int, 3> GPUContext::GetCUDAMaxGridDimSize() const {
  return impl_->max_grid_dim_size_;
}

Eigen::GpuDevice* GPUContext::eigen_device() const {
  return impl_->eigen_device();
}

W
Wilber 已提交
839
DnnWorkspaceHandle GPUContext::cudnn_workspace_handle() const {
W
Wilber 已提交
840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890
  return impl_->GetDnnWorkspace();
}

void GPUContext::CublasCall(
    const std::function<void(blasHandle_t)>& callback) const {
  impl_->CublasCall(callback);
}

void GPUContext::TensorCoreCublasCallIfAvailable(
    const std::function<void(blasHandle_t)>& callback) const {
  impl_->TensorCoreCublasCallIfAvailable(callback);
}

void GPUContext::CusparseCall(
    const std::function<void(sparseHandle_t)>& callback) const {
  impl_->CusparseCall(callback);
}

void GPUContext::RecordEvent(gpuEvent_t ev,
                             const std::function<void()>& callback) const {
  impl_->RecordEvent(ev, callback);
}

void GPUContext::RecordEvent(gpuEvent_t ev) const { impl_->RecordEvent(ev); }

void GPUContext::AddStreamCallback(
    const std::function<void()>& callback) const {
  impl_->AddStreamCallback(callback);
}

void GPUContext::WaitStreamCallback() const { impl_->WaitStreamCallback(); }

ncclComm_t GPUContext::nccl_comm() const { return impl_->GetNcclComm(); }

void GPUContext::set_nccl_comm(ncclComm_t comm) { impl_->SetNcclComm(comm); }

void GPUContext::Init() {
  impl_->allocator_ = const_cast<Allocator*>(&this->GetAllocator());
  impl_->Init();
}

void GPUContext::SetStream(gpuStream_t stream) { impl_->SetStream(stream); }

void GPUContext::SetEigenDevice(Eigen::GpuDevice* device) {
  impl_->SetEigenDevice(device);
}

void GPUContext::SetBlasHandle(blasHandle_t blas) {
  impl_->SetBlasHandle(blas);
}

891 892 893 894
void GPUContext::SetBlasLtHandle(blasLtHandle_t blaslt) {
  impl_->SetBlasLtHandle(blaslt);
}

W
Wilber 已提交
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941
void GPUContext::SetDnnHandle(dnnHandle_t handle) {
  impl_->SetDnnHandle(handle);
}

void GPUContext::SetSolverHandle(solverHandle_t handle) {
  impl_->SetSolverHandle(handle);
}

void GPUContext::SetSparseHandle(sparseHandle_t handle) {
  impl_->SetSparseHandle(handle);
}

void GPUContext::SetDnnWorkspaceHandle(DnnWorkspaceHandle* handle) {
  impl_->workspace_ = handle;
}

void GPUContext::PartialInitWithoutAllocator() {
  impl_->PartialInitWithoutAllocator();
}

void GPUContext::PartialInitWithAllocator() {
  impl_->allocator_ = const_cast<Allocator*>(&this->GetAllocator());
  impl_->PartialInitWithAllocator();
}

void GPUContext::SetComputeCapability(int val) {
  impl_->compute_capability_ = val;
}

void GPUContext::SetMaxThreadsPerMultiProcessor(int val) {
  impl_->max_threads_per_mp_ = val;
}

void GPUContext::SetMultiProcessors(int val) { impl_->multi_process_ = val; }

void GPUContext::SetMaxThreadsPerBlock(int val) {
  impl_->max_threads_per_block_ = val;
}

void GPUContext::SetMaxGridDimSize(const std::array<int, 3>& val) {
  impl_->max_grid_dim_size_ = val;
}

void GPUContext::SetDriverVersion(int val) { impl_->driver_version_ = val; }

void GPUContext::SetRuntimeVersion(int val) { impl_->runtime_version_ = val; }

942
}  // namespace phi