c_allreduce_op.h 13.3 KB
Newer Older
1
/* Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
2 3 4 5 6 7 8 9 10 11 12 13 14 15

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 <string>
18 19 20 21

#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/op_registry.h"
22 23
#include "paddle/fluid/memory/memcpy.h"
#include "paddle/fluid/memory/memory.h"
24
#include "paddle/fluid/operators/npu_op_runner.h"
25

26
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL) || \
27
    defined(PADDLE_WITH_ASCEND_CL) || defined(PADDLE_WITH_XPU_BKCL)
28
#include "paddle/fluid/platform/collective_helper.h"
29 30 31
#endif

#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
32 33 34
#include "paddle/fluid/platform/nccl_helper.h"
#endif

35 36 37 38
#if defined(PADDLE_WITH_XPU_BKCL)
#include "paddle/fluid/platform/bkcl_helper.h"
#endif

39 40 41 42 43
#if defined(PADDLE_WITH_GLOO)
#include <gloo/allreduce.h>
#include "paddle/fluid/framework/fleet/gloo_wrapper.h"
#endif

44 45 46 47
#if defined(PADDLE_WITH_ASCEND_CL)
#include "paddle/fluid/platform/hccl_helper.h"
#endif

48 49 50
namespace paddle {
namespace operators {

51 52 53 54 55 56 57 58 59 60 61 62 63
enum ReduceType { kRedSum, kRedMax, kRedMin, kRedProd };

class CAllReduceOp : public framework::OperatorWithKernel {
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

  void InferShape(framework::InferShapeContext* ctx) const override {
    ctx->SetOutputDim("Out", ctx->GetInputDim("X"));
  }

 protected:
  framework::OpKernelType GetExpectedKernelType(
      const framework::ExecutionContext& ctx) const override {
64 65
    return framework::OpKernelType(
        OperatorWithKernel::IndicateVarDataType(ctx, "X"), ctx.GetPlace());
66 67 68 69 70 71 72
  }
};

template <ReduceType red_type, typename T>
class CAllReduceOpCPUKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
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
#if defined(PADDLE_WITH_GLOO)
    auto in = ctx.Input<framework::Tensor>("X");
    auto out = ctx.Output<framework::Tensor>("Out");

    auto place = ctx.GetPlace();
    int64_t send_numel = in->numel();
    const T* send_buff = in->data<T>();
    T* recv_buff = out->mutable_data<T>(in->dims(), place);
    auto gloo = paddle::framework::GlooWrapper::GetInstance();
    PADDLE_ENFORCE_EQ(
        gloo->IsInitialized(), true,
        platform::errors::PreconditionNotMet(
            "You must initialize the gloo environment first to use it."));
    gloo::AllreduceOptions opts(gloo->GetContext());
    opts.setInput(const_cast<T*>(send_buff), send_numel);
    opts.setOutput(recv_buff, send_numel);
    switch (red_type) {
      case kRedSum:
        opts.setReduceFunction(
            static_cast<void (*)(void*, const void*, const void*, size_t)>(
                &gloo::sum<T>));
        break;
      case kRedMax:
        opts.setReduceFunction(
            static_cast<void (*)(void*, const void*, const void*, size_t)>(
                &gloo::max<T>));
        break;
      case kRedMin:
        opts.setReduceFunction(
            static_cast<void (*)(void*, const void*, const void*, size_t)>(
                &gloo::min<T>));
        break;
      case kRedProd:
        opts.setReduceFunction(
            static_cast<void (*)(void*, const void*, const void*, size_t)>(
                &gloo::product<T>));
        break;
      default:
        PADDLE_ENFORCE_EQ(true, false,
                          platform::errors::InvalidArgument(
                              "Invalid reduce type: %d.", red_type));
    }
    gloo::allreduce(opts);
#else
    PADDLE_THROW(platform::errors::Unavailable(
        "PaddlePaddle should compile with GLOO by setting WITH_GLOO=ON"));
#endif
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
#if defined(PADDLE_WITH_ASCEND_CL)
// return true if found_inf_or_nan or return false;
template <typename T>
bool CheckNumerics(const framework::ExecutionContext& exe_ctx,
                   aclrtStream stream, const paddle::framework::Tensor* in) {
  auto& dev_ctx =
      exe_ctx.template device_context<paddle::platform::NPUDeviceContext>();
  using Tensor = paddle::framework::Tensor;
  Tensor out(in->type());
  out.Resize(in->dims());
  out.mutable_data<T>(dev_ctx.GetPlace());

  bool found_inf_data = false;

  try {
    const auto& runner =
        NpuOpRunner("CheckNumerics", {*in}, {out},
                    {{"message", std::string("check_numberics")}});
    runner.Run(stream);
    dev_ctx.Wait();
  } catch (platform::EnforceNotMet& exception) {
    LOG(WARNING) << "[check_nan_and_inf] detected contains NaN or INF!!!";
    found_inf_data = true;
  } catch (...) {
    LOG(WARNING) << "[check_nan_and_inf] detected contains NaN or INF!!!";
    found_inf_data = true;
  }

  return found_inf_data;
}
#endif

155 156 157 158 159
template <ReduceType red_type, typename T>
class CAllReduceOpASCENDKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
#if defined(PADDLE_WITH_ASCEND_CL)
160 161
    auto in = ctx.Input<framework::Tensor>("X");
    auto out = ctx.Output<framework::Tensor>("Out");
162 163 164 165 166
    auto place = ctx.GetPlace();
    HcclDataType dtype = platform::ToHCCLDataType(in->type());
    int64_t numel = in->numel();

    void* sendbuff = reinterpret_cast<void*>(const_cast<T*>(in->data<T>()));
167
    out->mutable_data<T>(in->dims(), ctx.GetPlace());
168 169 170 171 172 173 174 175 176
    void* recvbuff = reinterpret_cast<void*>(out->data<T>());

    int ring_id = ctx.Attr<int>("ring_id");
    std::string group =
        std::string(HCOM_GROUP_PREFIX) + std::to_string(ring_id);
    auto comm =
        paddle::platform::HCCLCommContext::Instance().Get(ring_id, place);

    aclrtStream stream = nullptr;
177 178
    auto dev_ctx = static_cast<platform::NPUDeviceContext*>(
        platform::DeviceContextPool::Instance().Get(place));
179
    if (ctx.Attr<bool>("use_calc_stream")) {
180
      stream = dev_ctx->stream();
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
    } else {
      stream = comm->stream();
    }

    HcclReduceOp hccl_red_type = HCCL_REDUCE_SUM;
    switch (red_type) {
      case kRedSum:
        hccl_red_type = HCCL_REDUCE_SUM;
        break;

      case kRedMax:
        hccl_red_type = HCCL_REDUCE_MAX;
        break;

      case kRedMin:
        hccl_red_type = HCCL_REDUCE_MIN;
        break;

      case kRedProd:
        hccl_red_type = HCCL_REDUCE_PROD;
        break;

      default:
        PADDLE_THROW(platform::errors::InvalidArgument(
            "Invalid reduce type: %d", red_type));
    }

208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243
    VLOG(3) << "hccl allreduce, parameter is: "
            << "input num: " << in->dims() << "dtype: " << dtype
            << "hccl_red_type: " << hccl_red_type << ", group is: " << group
            << ", sendbuff:" << sendbuff << ", recvbuff:" << recvbuff
            << ", out_size:" << out->memory_size()
            << ", use_calc_stream:" << ctx.Attr<bool>("use_calc_stream")
            << ", stream:" << stream;

    framework::Tensor tmp;
    tmp.mutable_data<float>({8}, ctx.GetPlace());

    bool check_numerics = false;

    auto d_type = in->type();
    switch (d_type) {
      case framework::proto::VarType::FP16:
      case framework::proto::VarType::FP32: {
        VLOG(4) << "prepare to FoundNanInf";
        check_numerics = CheckNumerics<T>(ctx, dev_ctx->stream(), in);
        VLOG(4) << "check_numerics:" << check_numerics;
        break;
      }
      default:
        break;
    }

    if (check_numerics) {
      T inf = static_cast<T>(std::numeric_limits<float>::infinity());
      VLOG(4) << "fill input data constant inf";
      auto dims = in->dims();
      auto mutable_in = const_cast<framework::Tensor*>(in);
      FillNpuTensorWithConstant<T>(mutable_in, inf);
      mutable_in->Resize(dims);
    }

    VLOG(3) << "hccl allreduce, parameter is: "
244
            << "input num: " << numel << "dtype: " << dtype
245 246 247
            << "hccl_red_type: " << hccl_red_type << ", group is: " << group
            << ", sendbuff:" << sendbuff << ", recvbuff:" << recvbuff
            << ", out_size:" << out->memory_size();
248 249 250 251 252 253 254 255 256 257 258 259 260

    PADDLE_ENFORCE_NPU_SUCCESS(platform::dynload::HcclAllReduce(
        sendbuff, recvbuff, numel, dtype, hccl_red_type, comm->comm(),
        reinterpret_cast<void*>(stream)));

    out->Resize(in->dims());
#else
    PADDLE_THROW(platform::errors::PreconditionNotMet(
        "PaddlePaddle should compile with NPU."));
#endif
  }
};

261 262 263 264 265 266 267 268 269 270 271
template <ReduceType red_type, typename T>
class CAllReduceOpXPUKernel : public framework::OpKernel<T> {
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
#if defined(PADDLE_WITH_XPU_BKCL)
    auto in = ctx.Input<framework::Tensor>("X");
    auto out = ctx.Output<framework::Tensor>("Out");

    auto place = ctx.GetPlace();
    BKCLDataType dtype = platform::ToBKCLDataType(in->type());
    int64_t numel = in->numel();
272
    const void* sendbuff = in->data<T>();
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
    out->Resize(in->dims());
    void* recvbuff = out->mutable_data<T>(place);

    int rid = ctx.Attr<int>("ring_id");
    auto comm = platform::BKCLCommContext::Instance().Get(rid, place);

    XPUStream stream = nullptr;
    if (ctx.Attr<bool>("use_calc_stream")) {
      auto dev_ctx = platform::DeviceContextPool::Instance().Get(place);
      stream = static_cast<platform::XPUDeviceContext*>(dev_ctx)
                   ->x_context()
                   ->xpu_stream;
    } else {
      stream = comm->stream();
    }

    BKCLOp bkcl_red_type = BKCL_ADD;
    switch (red_type) {
      case kRedSum:
        bkcl_red_type = BKCL_ADD;
        break;

      case kRedMax:
        bkcl_red_type = BKCL_MAX;
        break;

      case kRedMin:
        bkcl_red_type = BKCL_MIN;
        break;

      case kRedProd:
        bkcl_red_type = BKCL_PRODUCT;
        break;

      default:
        PADDLE_THROW(platform::errors::InvalidArgument(
            "Invalid reduce type: %d", red_type));
    }

    PADDLE_ENFORCE_EQ(bkcl_all_reduce(comm->comm(), sendbuff, recvbuff, numel,
                                      dtype, bkcl_red_type, stream),
                      BKCL_SUCCESS, platform::errors::PreconditionNotMet(
                                        "BKCL all reduce failed"));
#else
    PADDLE_THROW(platform::errors::PreconditionNotMet(
        "PaddlePaddle should be compiled with XPU."));
#endif
  }
};

323 324
template <ReduceType red_type, typename T>
class CAllReduceOpCUDAKernel : public framework::OpKernel<T> {
325 326
 public:
  void Compute(const framework::ExecutionContext& ctx) const override {
327
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
328 329 330
    auto in = ctx.Input<framework::Tensor>("X");
    auto out = ctx.Output<framework::Tensor>("Out");

331
    auto place = ctx.GetPlace();
332 333
    ncclDataType_t dtype = platform::ToNCCLDataType(in->type());
    int64_t numel = in->numel();
334
    const void* sendbuff = in->data<T>();
335 336 337 338
    out->Resize(in->dims());
    void* recvbuff = out->mutable_data<T>(place);

    int rid = ctx.Attr<int>("ring_id");
339
    auto comm = platform::NCCLCommContext::Instance().Get(rid, place);
340

341
    gpuStream_t stream = nullptr;
342 343 344 345 346 347 348
    if (ctx.Attr<bool>("use_calc_stream")) {
      auto dev_ctx = platform::DeviceContextPool::Instance().Get(place);
      stream = static_cast<platform::CUDADeviceContext*>(dev_ctx)->stream();
    } else {
      stream = comm->stream();
    }

349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
    ncclRedOp_t nccl_red_type = ncclSum;
    switch (red_type) {
      case kRedSum:
        nccl_red_type = ncclSum;
        break;

      case kRedMax:
        nccl_red_type = ncclMax;
        break;

      case kRedMin:
        nccl_red_type = ncclMin;
        break;

      case kRedProd:
        nccl_red_type = ncclProd;
        break;

      default:
M
MRXLT 已提交
368 369
        PADDLE_THROW(platform::errors::InvalidArgument(
            "Invalid reduce type: %d", red_type));
370 371
    }

372
    PADDLE_ENFORCE_CUDA_SUCCESS(platform::dynload::ncclAllReduce(
373
        sendbuff, recvbuff, numel, dtype, nccl_red_type, comm->comm(), stream));
374
#else
M
MRXLT 已提交
375 376
    PADDLE_THROW(platform::errors::PreconditionNotMet(
        "PaddlePaddle should compile with GPU."));
377 378 379 380
#endif
  }
};

381 382 383 384 385 386 387
class CAllReduceOpMaker : public framework::OpProtoAndCheckerMaker {
 public:
  void Make() {
    AddInput("X", "(Tensor), tensor to be allreduced.");
    AddOutput("Out", "(Tensor) the allreduced result.");
    AddAttr<int>("ring_id", "(int default 0) communication ring id.")
        .SetDefault(0);
388 389 390 391
#if defined(PADDLE_WITH_ASCEND_CL)
    AddAttr<std::string>("tag", "(string default tag) tag for all reduce.")
        .SetDefault("tag");
#endif
392 393 394 395
    AddAttr<bool>(
        "use_calc_stream",
        "(bool default false) eject CUDA operations to calculation stream.")
        .SetDefault(false);
L
lilong12 已提交
396 397 398 399 400 401
    AddAttr<bool>(
        "use_model_parallel",
        "(bool default false) use this op with model parallel mode. In model "
        "parallel mode, the backward is c_identity which returns itself for "
        "c_allreduce_sum.")
        .SetDefault(false);
402 403 404 405 406 407 408 409 410 411 412 413 414 415
    AddComment(string::Sprintf(R"DOC(
CAllReduce %s Operator

Call collective AllReduce with reduce type %s. If input and output are
the same variable, in-place allreduce will be used.
Reference: https://docs.nvidia.com/deeplearning/sdk/nccl-developer-guide/docs/usage/operations.html#allreduce
)DOC",
                               GetName(), GetName()));
  }

 protected:
  virtual std::string GetName() const = 0;
};

416 417
}  // namespace operators
}  // namespace paddle