| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/gpu/nccl/ |
| D | nccl_collective_gpu_kernel.cc | 42 …ReduceScatter, KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeFloat32).AddOutputAttr(kN…
45 …ReduceScatter, KernelAttr().AddAllSameAttr(true).AddInputAttr(kNumberTypeFloat16).AddOutputAttr(kN…
47 MS_REG_GPU_KERNEL_ONE(ReduceScatter,
|
| D | nccl_gpu_kernel.h | 37 typedef ncclResult_t (*ReduceScatter)(const void *, void *, size_t, ncclDataType_t, ncclRedOp_t, cu… typedef
|
| D | nccl_collective_gpu_kernel.h | 188 … reinterpret_cast<ReduceScatter>(dlsym(const_cast<void *>(collective_handle_), "ReduceScatter")); in LaunchReduceScatter()
|
| /third_party/mindspore/tests/st/nccl/ |
| D | test_nccl_reduce_scatter_op.py | 43 self.reduce_scatter1 = P.ReduceScatter(self.op0, group=NCCL_WORLD_COMM_GROUP)
44 self.reduce_scatter2 = P.ReduceScatter(self.op1, group=NCCL_WORLD_COMM_GROUP)
45 self.reduce_scatter3 = P.ReduceScatter(self.op2, group=NCCL_WORLD_COMM_GROUP)
|
| /third_party/mindspore/mindspore/ccsrc/runtime/device/gpu/distribution/ |
| D | collective_wrapper.cc | 47 ncclResult_t ReduceScatter(const void *input_addr, void *output_addr, size_t count, ncclDataType_t … in ReduceScatter() function
49 …return NCCLWrapper::instance().ReduceScatter(input_addr, output_addr, count, data_type, reduce_typ… in ReduceScatter()
|
| D | nccl_wrapper.h | 42 …ncclResult_t ReduceScatter(const void *input_addr, void *output_addr, size_t count, ncclDataType_t…
|
| D | collective_wrapper.h | 45 extern "C" EXPORT_WRAPPER ncclResult_t ReduceScatter(const void *input_addr, void *output_addr, siz…
|
| D | nccl_wrapper.cc | 65 ncclResult_t NCCLWrapper::ReduceScatter(const void *input_addr, void *output_addr, size_t count, in ReduceScatter() function in mindspore::device::gpu::NCCLWrapper
|
| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/hccl/ |
| D | hcom_all_reduce_scatter.h | 38 MS_HCCL_REG_KERNEL(ReduceScatter, HcomAllReduceScatterKernel);
|
| /third_party/mindspore/mindspore/ccsrc/runtime/device/cpu/mpi/ |
| D | mpi_adapter.h | 38 …FUNC_EXPORT bool ReduceScatter(const float *input, float *output, const std::vector<int> &ranks_gr…
|
| D | mpi_export.cc | 43 return inst->ReduceScatter(input, output, ranks_group, data_num, op_type); in MPIReduceScatter()
|
| D | mpi_adapter.cc | 163 bool MPIAdapter::ReduceScatter(const float *input, float *output, const std::vector<int> &ranks_gro… in ReduceScatter() function in mindspore::device::cpu::MPIAdapter
|
| /third_party/mindspore/tests/ut/python/communication/ |
| D | test_comm.py | 28 …om mindspore.ops.operations.comm_ops import AllReduce, AllGather, AlltoAll, ReduceOp, ReduceScatter
101 self.reducescatter = ReduceScatter(op)
|
| /third_party/mindspore/mindspore/ops/_grad/ |
| D | grad_comm_ops.py | 27 … ReduceScatter, _HostReduceScatter, _VirtualDiv, _VirtualAdd, AllSwap,
217 reduce_scatter = ReduceScatter(ReduceOp.SUM, self.group).add_prim_attr("fusion", fusion)
309 @bprop_getters.register(ReduceScatter)
|
| /third_party/mindspore/mindspore/ops/operations/ |
| D | __init__.py | 38 from .comm_ops import (AllGather, AllReduce, NeighborExchange, AlltoAll, AllSwap, ReduceScatter, Br…
|
| D | comm_ops.py | 367 class ReduceScatter(PrimitiveWithInfer): class
|
| /third_party/mindspore/ |
| D | RELEASE.md | 712 - [STABLE] Support AllGather and ReduceScatter fusion.(Ascend)
|