Bertan Dogancay
96 lines
3.7 KiB
C++
96 lines
3.7 KiB
C++
/*************************************************************************
|
|
* Copyright (c) Microsoft Corporation.
|
|
* Licensed under the MIT License.
|
|
************************************************************************/
|
|
|
|
#include "enqueue.h"
|
|
#include "msccl/msccl_parser.h"
|
|
#include "msccl/msccl_setup.h"
|
|
#include "msccl/msccl_status.h"
|
|
#include "api_trace.h"
|
|
#include <cstdio>
|
|
#include <cstdlib>
|
|
|
|
NCCL_API(ncclResult_t, mscclLoadAlgo, const char *mscclAlgoFilePath, mscclAlgoHandle_t *mscclAlgoHandle, int rank);
|
|
ncclResult_t mscclLoadAlgo_impl(const char *mscclAlgoFilePath, mscclAlgoHandle_t *mscclAlgoHandle, int rank) {
|
|
mscclStatus& status = mscclGetStatus(rank);
|
|
|
|
if (status.freeAlgoHandles.size() == 0) {
|
|
WARN("MSCCL: MSCCL_MAX_NUM_ALGOS (%d) limit reached", MSCCL_MAX_NUM_ALGOS);
|
|
return ncclInvalidUsage;
|
|
}
|
|
*mscclAlgoHandle = *status.freeAlgoHandles.rbegin();
|
|
status.freeAlgoHandles.pop_back();
|
|
|
|
struct mscclAlgo* hostAlgo;
|
|
NCCLCHECK(ncclCalloc(&hostAlgo, 1));
|
|
NCCLCHECK(mscclGetAlgoFromXmlFile(mscclAlgoFilePath, hostAlgo, rank));
|
|
status.hostAlgos[*mscclAlgoHandle] = hostAlgo;
|
|
|
|
struct mscclAlgo* devAlgo;
|
|
NCCLCHECK(ncclCudaMalloc(&devAlgo, 1));
|
|
CUDACHECK(hipMemcpy(devAlgo, hostAlgo, sizeof(struct mscclAlgo), hipMemcpyHostToDevice));
|
|
status.devAlgos[*mscclAlgoHandle] = devAlgo;
|
|
|
|
return ncclSuccess;
|
|
}
|
|
|
|
NCCL_API(ncclResult_t, mscclRunAlgo,
|
|
const void* sendBuff, const size_t sendCounts[], const size_t sDisPls[],
|
|
void* recvBuff, const size_t recvCounts[], const size_t rDisPls[],
|
|
size_t count, ncclDataType_t dataType, int root, int peer, ncclRedOp_t op,
|
|
mscclAlgoHandle_t mscclAlgoHandle, ncclComm_t comm, hipStream_t stream);
|
|
ncclResult_t mscclRunAlgo_impl(
|
|
const void* sendBuff, const size_t sendCounts[], const size_t sDisPls[],
|
|
void* recvBuff, const size_t recvCounts[], const size_t rDisPls[],
|
|
size_t count, ncclDataType_t dataType, int root, int peer, ncclRedOp_t op,
|
|
mscclAlgoHandle_t mscclAlgoHandle, ncclComm_t comm, hipStream_t stream) {
|
|
struct NvtxParamsMsccl {
|
|
size_t bytes;
|
|
ncclRedOp_t op;
|
|
ncclDataType_t dataType;
|
|
};
|
|
// Just pass the size of one send/recv messages and not the total bytes sent/received.
|
|
constexpr nvtxPayloadSchemaEntry_t MscclSchema[] = {
|
|
{0, NVTX_PAYLOAD_ENTRY_TYPE_SIZE, "Message size [bytes]"},
|
|
{0, NVTX_PAYLOAD_ENTRY_NCCL_REDOP, "Reduction operation", nullptr, 0,
|
|
offsetof(NvtxParamsMsccl, op)},
|
|
{0, NVTX_PAYLOAD_ENTRY_TYPE_DATATYPE, "Data type", nullptr, 0,
|
|
offsetof(NvtxParamsMsccl, dataType)}
|
|
};
|
|
NvtxParamsMsccl payload{count * ncclTypeSize(dataType), op, dataType};
|
|
NVTX3_FUNC_WITH_PARAMS(MSCCL, MscclSchema, payload)
|
|
|
|
mscclStatus& status = mscclGetStatus(comm->rank);
|
|
struct mscclAlgo* hostAlgo = status.hostAlgos[mscclAlgoHandle];
|
|
struct mscclAlgo* devAlgo = status.devAlgos[mscclAlgoHandle];
|
|
|
|
// NCCL adds a lot of guarantees that target device is getting used
|
|
// in its group management code, which we entirely skip when MSCCL is used
|
|
// Therefore, in single thread multiGPU mode
|
|
// setting the device is critical to be sure
|
|
// communication is done on the intended device
|
|
|
|
CUDACHECK(hipSetDevice(comm->cudaDev));
|
|
|
|
NCCLCHECK(mscclGetCaptureStatus(comm->rank, stream));
|
|
|
|
NCCLCHECK(mscclSetupCount(hostAlgo, comm, count, dataType));
|
|
|
|
NCCLCHECK(mscclSetupScratch(hostAlgo, stream));
|
|
|
|
NCCLCHECK(mscclSetupSyncFlags(comm->rank, stream));
|
|
|
|
NCCLCHECK(mscclSetupProxy(hostAlgo, comm, stream));
|
|
|
|
NCCLCHECK(mscclSetupKernel(sendBuff, recvBuff, count, dataType, op, hostAlgo, devAlgo, comm, stream));
|
|
|
|
return ncclSuccess;
|
|
}
|
|
|
|
NCCL_API(ncclResult_t, mscclUnloadAlgo, mscclAlgoHandle_t mscclAlgoHandle);
|
|
ncclResult_t mscclUnloadAlgo_impl(mscclAlgoHandle_t mscclAlgoHandle) {
|
|
// deprecated
|
|
return ncclSuccess;
|
|
}
|