enforce.h 28.9 KB
Newer Older
1
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

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

17 18 19 20
#ifdef __GNUC__
#include <cxxabi.h>  // for __cxa_demangle
#endif               // __GNUC__

21
#if !defined(_WIN32)
22
#include <dlfcn.h>   // dladdr
23
#include <unistd.h>  // sleep, usleep
24
#else                // _WIN32
25 26 27
#ifndef NOMINMAX
#define NOMINMAX  // msvc max/min macro conflict with std::min/max
#endif
28
#include <windows.h>  // GetModuleFileName, Sleep
29 30
#endif

31 32 33
#ifdef PADDLE_WITH_CUDA
#include <cublas_v2.h>
#include <cudnn.h>
34
#include <cufft.h>
35
#include <curand.h>
Z
zhangkaihuo 已提交
36
#include <cusparse.h>
37 38
#include <thrust/system/cuda/error.h>
#include <thrust/system_error.h>
39
#include "paddle/fluid/platform/external_error.pb.h"
40 41
#endif  // PADDLE_WITH_CUDA

42 43 44 45 46
#ifdef PADDLE_WITH_HIP
#include <hiprand.h>
#include <miopen/miopen.h>
#include <rocblas.h>
#include <thrust/system/hip/error.h>
47
#include <thrust/system_error.h>  // NOLINT
48 49
#endif

50
#include <fstream>
Y
Yu Yang 已提交
51
#include <iomanip>
L
liaogang 已提交
52
#include <memory>
53 54 55
#include <sstream>
#include <stdexcept>
#include <string>
S
sneaxiy 已提交
56 57
#include <type_traits>
#include <utility>
58

chen.zhiyu's avatar
chen.zhiyu 已提交
59 60 61 62
#if !defined(_WIN32) && !defined(PADDLE_WITH_MUSL)
#include <execinfo.h>
#endif

63
#define GLOG_NO_ABBREVIATED_SEVERITIES  // msvc conflict logging with windows.h
64
#include "gflags/gflags.h"
65
#include "glog/logging.h"
66
#include "paddle/fluid/platform/errors.h"
Y
Yi Wang 已提交
67
#include "paddle/fluid/platform/macros.h"
68
#include "paddle/fluid/platform/variant.h"
69 70
#include "paddle/fluid/string/printf.h"
#include "paddle/fluid/string/to_string.h"
71
#include "paddle/pten/backends/dynload/port.h"
72

73
#ifdef PADDLE_WITH_CUDA
74 75 76 77
#include "paddle/pten/backends/dynload/cublas.h"
#include "paddle/pten/backends/dynload/cudnn.h"
#include "paddle/pten/backends/dynload/curand.h"
#include "paddle/pten/backends/dynload/cusolver.h"
78
#if !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
L
lilong12 已提交
79
#include <error.h>
80
#include "paddle/pten/backends/dynload/nccl.h"
Y
Yi Wang 已提交
81 82
#endif  // __APPLE__
#endif  // PADDLE_WITH_CUDA
83

84
#ifdef PADDLE_WITH_HIP
85 86 87 88
#include "paddle/pten/backends/dynload/hipfft.h"
#include "paddle/pten/backends/dynload/hiprand.h"
#include "paddle/pten/backends/dynload/miopen.h"
#include "paddle/pten/backends/dynload/rocblas.h"
89 90
#if !defined(__APPLE__) && defined(PADDLE_WITH_RCCL)
#include <error.h>  // NOLINT
91
#include "paddle/pten/backends/dynload/rccl.h"
92 93 94
#endif  // __APPLE__
#endif  // PADDLE_WITH_HIP

95 96 97
// Note: these headers for simplify demangle type string
#include "paddle/fluid/framework/type_defs.h"
#include "paddle/fluid/imperative/type_defs.h"
98
#include "paddle/pten/core/enforce.h"
99 100
// Note: this header for simplify HIP and CUDA type string
#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
101
#include "paddle/fluid/platform/device/gpu/gpu_types.h"
102
#endif
Z
Zeng Jinle 已提交
103
#include "paddle/fluid/platform/flags.h"
W
wanghuancoder 已提交
104

105
namespace pten {
W
wanghuancoder 已提交
106
class ErrorSummary;
107
}  // namespace pten
108

109
#if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP)
110 111
DECLARE_int64(gpu_allocator_retry_time);
#endif
112 113
DECLARE_int32(call_stack_level);

114 115
namespace paddle {
namespace platform {
116
using namespace ::pten::enforce;  // NOLINT
117

118 119
/** HELPER MACROS AND FUNCTIONS **/

Z
Zeng Jinle 已提交
120 121 122 123
#ifndef PADDLE_MAY_THROW
#define PADDLE_MAY_THROW noexcept(false)
#endif

124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
/*
 * Summary: This BOOST_GET(_**) series macros are used to call boost::get
 *   safely. boost::get is not a completely safe api, although it will not
 *   go wrong in most cases, but in extreme cases, it may fail and directly
 *   throw a boost::bad_get exception, without any stack information.
 *   This kind of problems is difficult to debug, so add these macros to
 *   enrich boost::get error information. At the same time, we restrict
 *   the direct use of boost::get by CI rule.
 *
 * Parameters:
 *     __TYPE: the target variable type
 *     __VALUE: the target variable to get
 *
 * Examples:
 *     - unsafe writing: int x = boost::get<int>(y);
 *     - safe writing: int x = BOOST_GET(int, y);
 *
 * Note: GCC 4.8 cannot select right overloaded function here, so need
 *    to define different functions and macros here, after we upgreade
 *    CI gcc version, we can only define one BOOST_GET macro.
144
 */
145 146
namespace details {

147 148
using namespace pten::enforce::details;  // NOLINT

149 150 151 152 153 154 155 156 157 158 159
#define DEFINE_SAFE_BOOST_GET(__InputType, __OutputType, __OutputTypePtr,      \
                              __FuncName)                                      \
  template <typename OutputType, typename InputType>                           \
  auto __FuncName(__InputType input, const char* expression, const char* file, \
                  int line)                                                    \
      ->typename std::conditional<std::is_pointer<InputType>::value,           \
                                  __OutputTypePtr, __OutputType>::type {       \
    try {                                                                      \
      return boost::get<OutputType>(input);                                    \
    } catch (boost::bad_get&) {                                                \
      HANDLE_THE_ERROR                                                         \
160 161
      throw ::pten::enforce::EnforceNotMet(                                    \
          pten::errors::InvalidArgument(                                       \
162 163
              "boost::get failed, cannot get value "                           \
              "(%s) by type %s, its type is %s.",                              \
164 165
              expression, pten::enforce::demangle(typeid(OutputType).name()),  \
              pten::enforce::demangle(input.type().name())),                   \
166 167 168 169 170 171 172 173 174 175 176 177 178
          file, line);                                                         \
      END_HANDLE_THE_ERROR                                                     \
    }                                                                          \
  }

DEFINE_SAFE_BOOST_GET(InputType&, OutputType&, OutputType*, SafeBoostGet);
DEFINE_SAFE_BOOST_GET(const InputType&, const OutputType&, const OutputType*,
                      SafeBoostGetConst);
DEFINE_SAFE_BOOST_GET(InputType&&, OutputType, OutputType*,
                      SafeBoostGetMutable);

}  // namespace details

179 180 181 182 183 184 185 186
#define BOOST_GET(__TYPE, __VALUE)                                             \
  paddle::platform::details::SafeBoostGet<__TYPE>(__VALUE, #__VALUE, __FILE__, \
                                                  __LINE__)
#define BOOST_GET_CONST(__TYPE, __VALUE)                                  \
  paddle::platform::details::SafeBoostGetConst<__TYPE>(__VALUE, #__VALUE, \
                                                       __FILE__, __LINE__)
#define BOOST_GET_MUTABLE(__TYPE, __VALUE)                                  \
  paddle::platform::details::SafeBoostGetMutable<__TYPE>(__VALUE, #__VALUE, \
187 188
                                                         __FILE__, __LINE__)

189 190
/** OTHER EXCEPTION AND ENFORCE **/

191 192
struct EOFException : public std::exception {
  std::string err_str_;
193
  EOFException(const char* err_msg, const char* file, int line) {
194
    err_str_ = paddle::string::Sprintf("%s at [%s:%d]", err_msg, file, line);
195 196
  }

197
  const char* what() const noexcept override { return err_str_.c_str(); }
198 199
};

200
#define PADDLE_THROW_EOF()                                                   \
201 202
  do {                                                                       \
    HANDLE_THE_ERROR                                                         \
203 204
    throw paddle::platform::EOFException("There is no next data.", __FILE__, \
                                         __LINE__);                          \
205
    END_HANDLE_THE_ERROR                                                     \
206
  } while (0)
M
minqiyang 已提交
207

208 209 210 211 212 213 214 215
#define PADDLE_THROW_BAD_ALLOC(...)                                       \
  do {                                                                    \
    HANDLE_THE_ERROR                                                      \
    throw ::paddle::memory::allocation::BadAlloc(                         \
        pten::ErrorSummary(__VA_ARGS__).to_string(), __FILE__, __LINE__); \
    END_HANDLE_THE_ERROR                                                  \
  } while (0)

216 217
/**************************************************************************/
/**************************** NVIDIA ERROR ********************************/
218
#ifdef PADDLE_WITH_CUDA
219

220
namespace details {
M
minqiyang 已提交
221

222 223 224 225 226 227 228 229 230 231 232
template <typename T>
struct ExternalApiType {};

#define DEFINE_EXTERNAL_API_TYPE(type, success_value, proto_type) \
  template <>                                                     \
  struct ExternalApiType<type> {                                  \
    using Type = type;                                            \
    static constexpr Type kSuccess = success_value;               \
    static constexpr const char* kTypeString = #proto_type;       \
    static constexpr platform::proto::ApiType kProtoType =        \
        platform::proto::ApiType::proto_type;                     \
233 234
  }

235 236 237 238
DEFINE_EXTERNAL_API_TYPE(cudaError_t, cudaSuccess, CUDA);
DEFINE_EXTERNAL_API_TYPE(curandStatus_t, CURAND_STATUS_SUCCESS, CURAND);
DEFINE_EXTERNAL_API_TYPE(cudnnStatus_t, CUDNN_STATUS_SUCCESS, CUDNN);
DEFINE_EXTERNAL_API_TYPE(cublasStatus_t, CUBLAS_STATUS_SUCCESS, CUBLAS);
Z
zhangkaihuo 已提交
239
DEFINE_EXTERNAL_API_TYPE(cusparseStatus_t, CUSPARSE_STATUS_SUCCESS, CUSPARSE);
240
DEFINE_EXTERNAL_API_TYPE(cusolverStatus_t, CUSOLVER_STATUS_SUCCESS, CUSOLVER);
241
DEFINE_EXTERNAL_API_TYPE(cufftResult_t, CUFFT_SUCCESS, CUFFT);
242
DEFINE_EXTERNAL_API_TYPE(CUresult, CUDA_SUCCESS, CU);
243 244 245

#if !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
DEFINE_EXTERNAL_API_TYPE(ncclResult_t, ncclSuccess, NCCL);
246
#endif
247 248 249 250 251 252 253 254 255 256

}  // namespace details

template <typename T>
inline const char* GetErrorMsgUrl(T status) {
  using __CUDA_STATUS_TYPE__ = decltype(status);
  platform::proto::ApiType proto_type =
      details::ExternalApiType<__CUDA_STATUS_TYPE__>::kProtoType;
  switch (proto_type) {
    case platform::proto::ApiType::CUDA:
257
    case platform::proto::ApiType::CU:
258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
      return "https://docs.nvidia.com/cuda/cuda-runtime-api/"
             "group__CUDART__TYPES.html#group__CUDART__TYPES_"
             "1g3f51e3575c2178246db0a94a430e0038";
      break;
    case platform::proto::ApiType::CURAND:
      return "https://docs.nvidia.com/cuda/curand/"
             "group__HOST.html#group__HOST_1gb94a31d5c165858c96b6c18b70644437";
      break;
    case platform::proto::ApiType::CUDNN:
      return "https://docs.nvidia.com/deeplearning/cudnn/api/"
             "index.html#cudnnStatus_t";
      break;
    case platform::proto::ApiType::CUBLAS:
      return "https://docs.nvidia.com/cuda/cublas/index.html#cublasstatus_t";
      break;
    case platform::proto::ApiType::CUSOLVER:
      return "https://docs.nvidia.com/cuda/cusolver/"
             "index.html#cuSolverSPstatus";
      break;
    case platform::proto::ApiType::NCCL:
      return "https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/api/"
             "types.html#ncclresult-t";
      break;
281 282
    case platform::proto::ApiType::CUFFT:
      return "https://docs.nvidia.com/cuda/cufft/index.html#cufftresult";
Z
zhangkaihuo 已提交
283 284 285 286
    case platform::proto::ApiType::CUSPARSE:
      return "https://docs.nvidia.com/cuda/cusparse/"
             "index.html#cusparseStatus_t";
      break;
287 288 289 290 291 292 293 294
    default:
      return "Unknown type of External API, can't get error message URL!";
      break;
  }
}

template <typename T>
inline std::string GetExternalErrorMsg(T status) {
295
  std::ostringstream sout;
296 297 298
  bool _initSucceed = false;
  platform::proto::ExternalErrorDesc externalError;
  if (externalError.ByteSizeLong() == 0) {
299 300 301
    std::string filePath;
#if !defined(_WIN32)
    Dl_info info;
302
    if (dladdr(reinterpret_cast<void*>(GetCurrentTraceBackString), &info)) {
303 304 305 306 307 308 309 310
      std::string strModule(info.dli_fname);
      const size_t last_slash_idx = strModule.find_last_of("/");
      std::string compare_path = strModule.substr(strModule.length() - 6);
      if (std::string::npos != last_slash_idx) {
        strModule.erase(last_slash_idx, std::string::npos);
      }
      if (compare_path.compare("avx.so") == 0) {
        filePath =
311 312 313 314 315
            strModule +
            "/../include/third_party/externalError/data/externalErrorMsg.pb";
      } else {
        filePath = strModule +
                   "/../../third_party/externalError/data/externalErrorMsg.pb";
316 317 318
      }
    }
#else
319
    char buf[512];
320 321
    MEMORY_BASIC_INFORMATION mbi;
    HMODULE h_module =
322
        (::VirtualQuery(GetCurrentTraceBackString, &mbi, sizeof(mbi)) != 0)
323 324
            ? (HMODULE)mbi.AllocationBase
            : NULL;
325
    GetModuleFileName(h_module, buf, 512);
326 327 328 329 330 331 332
    std::string strModule(buf);
    const size_t last_slash_idx = strModule.find_last_of("\\");
    std::string compare_path = strModule.substr(strModule.length() - 7);
    if (std::string::npos != last_slash_idx) {
      strModule.erase(last_slash_idx, std::string::npos);
    }
    if (compare_path.compare("avx.pyd") == 0) {
333 334 335
      filePath = strModule +
                 "\\..\\include\\third_"
                 "party\\externalerror\\data\\externalErrorMsg.pb";
336 337
    } else {
      filePath =
338 339
          strModule +
          "\\..\\..\\third_party\\externalerror\\data\\externalErrorMsg.pb";
340 341 342
    }
#endif
    std::ifstream fin(filePath, std::ios::in | std::ios::binary);
343
    _initSucceed = externalError.ParseFromIstream(&fin);
344
  }
345 346 347
  using __CUDA_STATUS_TYPE__ = decltype(status);
  platform::proto::ApiType proto_type =
      details::ExternalApiType<__CUDA_STATUS_TYPE__>::kProtoType;
348
  if (_initSucceed) {
349 350 351 352 353 354
    for (int i = 0; i < externalError.errors_size(); ++i) {
      if (proto_type == externalError.errors(i).type()) {
        for (int j = 0; j < externalError.errors(i).messages_size(); ++j) {
          if (status == externalError.errors(i).messages(j).code()) {
            sout << "\n  [Hint: "
                 << externalError.errors(i).messages(j).message() << "]";
355 356 357 358 359 360
            return sout.str();
          }
        }
      }
    }
  }
361 362 363 364 365 366

  sout << "\n  [Hint: Please search for the error code(" << status
       << ") on website (" << GetErrorMsgUrl(status)
       << ") to get Nvidia's official solution and advice about "
       << details::ExternalApiType<__CUDA_STATUS_TYPE__>::kTypeString
       << " Error.]";
367
  return sout.str();
368 369
}

370 371 372 373
template std::string GetExternalErrorMsg<cudaError_t>(cudaError_t);
template std::string GetExternalErrorMsg<curandStatus_t>(curandStatus_t);
template std::string GetExternalErrorMsg<cudnnStatus_t>(cudnnStatus_t);
template std::string GetExternalErrorMsg<cublasStatus_t>(cublasStatus_t);
Z
zhangkaihuo 已提交
374
template std::string GetExternalErrorMsg<cusparseStatus_t>(cusparseStatus_t);
375
template std::string GetExternalErrorMsg<cusolverStatus_t>(cusolverStatus_t);
376
template std::string GetExternalErrorMsg<cufftResult_t>(cufftResult_t);
377
template std::string GetExternalErrorMsg<CUresult>(CUresult);
378 379 380 381 382 383 384 385 386 387 388 389
#if !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
template std::string GetExternalErrorMsg<ncclResult_t>(ncclResult_t);
#endif

/*************** CUDA ERROR ***************/
inline bool is_error(cudaError_t e) { return e != cudaSuccess; }

inline std::string build_nvidia_error_msg(cudaError_t e) {
  std::ostringstream sout;
  sout << "CUDA error(" << e << "), " << cudaGetErrorString(e) << ". "
       << GetExternalErrorMsg(e);
  return sout.str();
390 391
}

392 393 394
/*************** CURAND ERROR ***************/
inline bool is_error(curandStatus_t stat) {
  return stat != CURAND_STATUS_SUCCESS;
395 396 397
}

inline std::string build_nvidia_error_msg(curandStatus_t stat) {
398 399 400
  std::ostringstream sout;
  sout << "CURAND error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
401 402
}

403
/*************** CUDNN ERROR ***************/
M
minqiyang 已提交
404 405
inline bool is_error(cudnnStatus_t stat) {
  return stat != CUDNN_STATUS_SUCCESS;
406 407
}

408
inline std::string build_nvidia_error_msg(cudnnStatus_t stat) {
409 410
  std::ostringstream sout;
  sout << "CUDNN error(" << stat << "), "
411
       << pten::dynload::cudnnGetErrorString(stat) << ". "
412 413
       << GetExternalErrorMsg(stat);
  return sout.str();
414 415
}

416
/*************** CUBLAS ERROR ***************/
M
minqiyang 已提交
417 418
inline bool is_error(cublasStatus_t stat) {
  return stat != CUBLAS_STATUS_SUCCESS;
419 420
}

421
inline std::string build_nvidia_error_msg(cublasStatus_t stat) {
422 423 424
  std::ostringstream sout;
  sout << "CUBLAS error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
425 426
}

Z
zhangkaihuo 已提交
427 428 429 430 431 432 433 434 435 436 437
/*************** CUSPARSE ERROR ***************/
inline bool is_error(cusparseStatus_t stat) {
  return stat != CUSPARSE_STATUS_SUCCESS;
}

inline std::string build_nvidia_error_msg(cusparseStatus_t stat) {
  std::ostringstream sout;
  sout << "CUSparse error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
}

438
/*************** CUSOLVER ERROR ***************/
G
Guo Sheng 已提交
439 440 441 442 443
inline bool is_error(cusolverStatus_t stat) {
  return stat != CUSOLVER_STATUS_SUCCESS;
}

inline std::string build_nvidia_error_msg(cusolverStatus_t stat) {
444 445 446
  std::ostringstream sout;
  sout << "CUSOLVER error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
G
Guo Sheng 已提交
447 448
}

449 450 451 452 453 454 455 456 457
/*************** CUFFT ERROR ***************/
inline bool is_error(cufftResult_t stat) { return stat != CUFFT_SUCCESS; }

inline std::string build_nvidia_error_msg(cufftResult_t stat) {
  std::ostringstream sout;
  sout << "CUFFT error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
}

458 459 460 461 462 463 464 465 466
/*************** CUresult ERROR ***************/
inline bool is_error(CUresult stat) { return stat != CUDA_SUCCESS; }

inline std::string build_nvidia_error_msg(CUresult stat) {
  std::ostringstream sout;
  sout << "CU error(" << stat << "). " << GetExternalErrorMsg(stat);
  return sout.str();
}

467
/**************** NCCL ERROR ****************/
468
#if !defined(__APPLE__) && defined(PADDLE_WITH_NCCL)
S
sneaxiy 已提交
469 470 471 472
inline bool is_error(ncclResult_t nccl_result) {
  return nccl_result != ncclSuccess;
}

473
inline std::string build_nvidia_error_msg(ncclResult_t nccl_result) {
474 475
  std::ostringstream sout;
  sout << "NCCL error(" << nccl_result << "), "
476
       << pten::dynload::ncclGetErrorString(nccl_result) << ". ";
L
lilong12 已提交
477 478 479 480 481 482 483 484 485
  if (errno == ENOSPC || errno == EAGAIN) {
    std::string detail(strerror(errno));
    detail += "\nPlease try one of the following solutions:";
    detail += "\n1. export NCCL_SHM_DISABLE=1;";
    detail += "\n2. export NCCL_P2P_LEVEL=SYS;";
    detail +=
        "\n3. Increase shared memory by setting the -shm-size "
        "option when starting docker container, e.g., setting "
        " -shm-size=2g.\n";
486
    sout << " Detail: " + detail;
L
lilong12 已提交
487
  }
488 489
  sout << GetExternalErrorMsg(nccl_result);
  return sout.str();
490
}
491
#endif  // not(__APPLE__) and PADDLE_WITH_NCCL
492

493
#define PADDLE_ENFORCE_GPU_SUCCESS(COND)                         \
494 495 496 497
  do {                                                           \
    auto __cond__ = (COND);                                      \
    using __CUDA_STATUS_TYPE__ = decltype(__cond__);             \
    constexpr auto __success_type__ =                            \
498
        ::paddle::platform::details::ExternalApiType<            \
499 500
            __CUDA_STATUS_TYPE__>::kSuccess;                     \
    if (UNLIKELY(__cond__ != __success_type__)) {                \
501
      auto __summary__ = pten::errors::External(                 \
502 503 504
          ::paddle::platform::build_nvidia_error_msg(__cond__)); \
      __THROW_ERROR_INTERNAL__(__summary__);                     \
    }                                                            \
505 506
  } while (0)

507 508 509 510 511 512 513 514 515 516
#define PADDLE_ENFORCE_CUDA_LAUNCH_SUCCESS(OP)                                 \
  do {                                                                         \
    auto res = cudaGetLastError();                                             \
    if (UNLIKELY(res != cudaSuccess)) {                                        \
      auto msg = ::paddle::platform::build_nvidia_error_msg(res);              \
      PADDLE_THROW(platform::errors::Fatal("CUDA error after kernel (%s): %s", \
                                           OP, msg));                          \
    }                                                                          \
  } while (0)

517
inline void retry_sleep(unsigned milliseconds) {
518
#ifdef _WIN32
519
  Sleep(milliseconds);
520
#else
521 522 523 524 525 526 527 528 529
  if (milliseconds < 1000) {
    // usleep argument must be less than 1,000,000. Reference:
    // https://pubs.opengroup.org/onlinepubs/7908799/xsh/usleep.html
    usleep(milliseconds * 1000);
  } else {
    // clip to sleep in seconds because we can not and don't have to
    // sleep for exact milliseconds
    sleep(milliseconds / 1000);
  }
530 531 532
#endif
}

533 534 535 536 537 538
#define PADDLE_RETRY_CUDA_SUCCESS(COND)                                 \
  do {                                                                  \
    auto __cond__ = (COND);                                             \
    int retry_count = 1;                                                \
    using __CUDA_STATUS_TYPE__ = decltype(__cond__);                    \
    constexpr auto __success_type__ =                                   \
539
        ::paddle::platform::details::ExternalApiType<                   \
540 541
            __CUDA_STATUS_TYPE__>::kSuccess;                            \
    while (UNLIKELY(__cond__ != __success_type__) && retry_count < 5) { \
542
      retry_sleep(FLAGS_gpu_allocator_retry_time);                      \
543 544 545 546
      __cond__ = (COND);                                                \
      ++retry_count;                                                    \
    }                                                                   \
    if (UNLIKELY(__cond__ != __success_type__)) {                       \
547
      auto __summary__ = pten::errors::External(                        \
548 549 550 551 552
          ::paddle::platform::build_nvidia_error_msg(__cond__));        \
      __THROW_ERROR_INTERNAL__(__summary__);                            \
    }                                                                   \
  } while (0)

553
#undef DEFINE_EXTERNAL_API_TYPE
554
#endif  // PADDLE_WITH_CUDA
S
add EQ  
Superjom 已提交
555

556 557
/**************************************************************************/
/***************************** HIP ERROR **********************************/
558 559 560 561 562 563 564 565 566 567 568
#ifdef PADDLE_WITH_HIP

/***** HIP ERROR *****/
inline bool is_error(hipError_t e) { return e != hipSuccess; }

inline std::string build_rocm_error_msg(hipError_t e) {
  std::ostringstream sout;
  sout << " Hip error(" << e << "), " << hipGetErrorString(e) << ".";
  return sout.str();
}

569
/***** HIPRAND ERROR *****/
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
inline bool is_error(hiprandStatus_t stat) {
  return stat != HIPRAND_STATUS_SUCCESS;
}

inline const char* hiprandGetErrorString(hiprandStatus_t stat) {
  switch (stat) {
    case HIPRAND_STATUS_SUCCESS:
      return "HIPRAND_STATUS_SUCCESS";
    case HIPRAND_STATUS_VERSION_MISMATCH:
      return "HIPRAND_STATUS_VERSION_MISMATCH";
    case HIPRAND_STATUS_NOT_INITIALIZED:
      return "HIPRAND_STATUS_NOT_INITIALIZED";
    case HIPRAND_STATUS_ALLOCATION_FAILED:
      return "HIPRAND_STATUS_ALLOCATION_FAILED";
    case HIPRAND_STATUS_TYPE_ERROR:
      return "HIPRAND_STATUS_TYPE_ERROR";
    case HIPRAND_STATUS_OUT_OF_RANGE:
      return "HIPRAND_STATUS_OUT_OF_RANGE";
    case HIPRAND_STATUS_LENGTH_NOT_MULTIPLE:
      return "HIPRAND_STATUS_LENGTH_NOT_MULTIPLE";
    case HIPRAND_STATUS_DOUBLE_PRECISION_REQUIRED:
      return "HIPRAND_STATUS_DOUBLE_PRECISION_REQUIRED";
    case HIPRAND_STATUS_LAUNCH_FAILURE:
      return "HIPRAND_STATUS_LAUNCH_FAILURE";
    case HIPRAND_STATUS_PREEXISTING_FAILURE:
      return "HIPRAND_STATUS_PREEXISTING_FAILURE";
    case HIPRAND_STATUS_INITIALIZATION_FAILED:
      return "HIPRAND_STATUS_INITIALIZATION_FAILED";
    case HIPRAND_STATUS_ARCH_MISMATCH:
      return "HIPRAND_STATUS_ARCH_MISMATCH";
    case HIPRAND_STATUS_INTERNAL_ERROR:
      return "HIPRAND_STATUS_INTERNAL_ERROR";
    case HIPRAND_STATUS_NOT_IMPLEMENTED:
      return "HIPRAND_STATUS_NOT_IMPLEMENTED";
    default:
      return "Unknown hiprand status";
  }
}

inline std::string build_rocm_error_msg(hiprandStatus_t stat) {
  std::string msg(" Hiprand error, ");
  return msg + hiprandGetErrorString(stat) + " ";
}

/***** MIOPEN ERROR *****/
inline bool is_error(miopenStatus_t stat) {
  return stat != miopenStatusSuccess;
}

inline std::string build_rocm_error_msg(miopenStatus_t stat) {
  std::string msg(" Miopen error, ");
621
  return msg + pten::dynload::miopenGetErrorString(stat) + " ";
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
}

/***** ROCBLAS ERROR *****/
inline bool is_error(rocblas_status stat) {
  return stat != rocblas_status_success;
}

inline const char* rocblasGetErrorString(rocblas_status stat) {
  switch (stat) {
    case rocblas_status_invalid_handle:
      return "rocblas_status_invalid_handle";
    case rocblas_status_memory_error:
      return "rocblas_status_memory_error";
    case rocblas_status_invalid_value:
      return "rocblas_status_invalid_value";
    case rocblas_status_not_implemented:
      return "rocblas_status_not_implemented";
    case rocblas_status_invalid_pointer:
      return "rocblas_status_invalid_pointer";
    case rocblas_status_invalid_size:
      return "rocblas_status_invalid_size";
    case rocblas_status_internal_error:
      return "rocblas_status_internal_error";
    default:
      return "Unknown cublas status";
  }
}

inline std::string build_rocm_error_msg(rocblas_status stat) {
  std::string msg(" Rocblas error, ");
  return msg + rocblasGetErrorString(stat) + " ";
}

/****** RCCL ERROR ******/
#if !defined(__APPLE__) && defined(PADDLE_WITH_RCCL)
inline bool is_error(ncclResult_t nccl_result) {
  return nccl_result != ncclSuccess;
}

inline std::string build_rocm_error_msg(ncclResult_t nccl_result) {
  std::string msg(" Rccl error, ");
663
  return msg + pten::dynload::ncclGetErrorString(nccl_result) + " ";
664 665 666
}
#endif  // not(__APPLE__) and PADDLE_WITH_NCCL

667 668 669 670 671
/***** HIPFFT ERROR *****/
inline bool is_error(hipfftResult_t stat) { return stat != HIPFFT_SUCCESS; }

inline std::string build_rocm_error_msg(hipfftResult_t stat) {
  std::string msg(" HIPFFT error, ");
672
  return msg + pten::dynload::hipfftGetErrorString(stat) + " ";
673 674
}

675 676 677
namespace details {

template <typename T>
678
struct ExternalApiType {};
679

680 681 682 683 684
#define DEFINE_EXTERNAL_API_TYPE(type, success_value) \
  template <>                                         \
  struct ExternalApiType<type> {                      \
    using Type = type;                                \
    static constexpr Type kSuccess = success_value;   \
685 686
  }

687 688 689 690
DEFINE_EXTERNAL_API_TYPE(hipError_t, hipSuccess);
DEFINE_EXTERNAL_API_TYPE(hiprandStatus_t, HIPRAND_STATUS_SUCCESS);
DEFINE_EXTERNAL_API_TYPE(miopenStatus_t, miopenStatusSuccess);
DEFINE_EXTERNAL_API_TYPE(rocblas_status, rocblas_status_success);
691
DEFINE_EXTERNAL_API_TYPE(hipfftResult_t, HIPFFT_SUCCESS);
692 693

#if !defined(__APPLE__) && defined(PADDLE_WITH_RCCL)
694
DEFINE_EXTERNAL_API_TYPE(ncclResult_t, ncclSuccess);
695 696 697 698
#endif

}  // namespace details

699
#define PADDLE_ENFORCE_GPU_SUCCESS(COND)                       \
700 701 702 703
  do {                                                         \
    auto __cond__ = (COND);                                    \
    using __CUDA_STATUS_TYPE__ = decltype(__cond__);           \
    constexpr auto __success_type__ =                          \
704
        ::paddle::platform::details::ExternalApiType<          \
705 706
            __CUDA_STATUS_TYPE__>::kSuccess;                   \
    if (UNLIKELY(__cond__ != __success_type__)) {              \
707
      auto __summary__ = pten::errors::External(               \
708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726
          ::paddle::platform::build_rocm_error_msg(__cond__)); \
      __THROW_ERROR_INTERNAL__(__summary__);                   \
    }                                                          \
  } while (0)

inline void retry_sleep(unsigned millisecond) {
#ifdef _WIN32
  Sleep(millisecond);
#else
  sleep(millisecond);
#endif
}

#define PADDLE_RETRY_CUDA_SUCCESS(COND)                                 \
  do {                                                                  \
    auto __cond__ = (COND);                                             \
    int retry_count = 1;                                                \
    using __CUDA_STATUS_TYPE__ = decltype(__cond__);                    \
    constexpr auto __success_type__ =                                   \
727
        ::paddle::platform::details::ExternalApiType<                   \
728 729 730 731 732 733 734
            __CUDA_STATUS_TYPE__>::kSuccess;                            \
    while (UNLIKELY(__cond__ != __success_type__) && retry_count < 5) { \
      retry_sleep(FLAGS_gpu_allocator_retry_time);                      \
      __cond__ = (COND);                                                \
      ++retry_count;                                                    \
    }                                                                   \
    if (UNLIKELY(__cond__ != __success_type__)) {                       \
735
      auto __summary__ = pten::errors::External(                        \
736 737 738 739 740
          ::paddle::platform::build_rocm_error_msg(__cond__));          \
      __THROW_ERROR_INTERNAL__(__summary__);                            \
    }                                                                   \
  } while (0)

741
#undef DEFINE_EXTERNAL_API_TYPE
742 743
#endif  // PADDLE_WITH_HIP

744 745
}  // namespace platform
}  // namespace paddle