| /external/tensorflow/tensorflow/core/nccl/ |
| D | collective_communicator.cc | 41 Status ReductionOp(const string& merge_op, ncclRedOp_t* reduction_op) { in ReductionOp() argument
43 *reduction_op = ncclSum; in ReductionOp()
46 *reduction_op = ncclProd; in ReductionOp()
49 *reduction_op = ncclMax; in ReductionOp()
52 *reduction_op = ncclMin; in ReductionOp()
137 ncclRedOp_t reduction_op; in Enqueue() local
139 ReductionOp(col_params->merge_op->type_string(), &reduction_op); in Enqueue()
145 reduction_op); in Enqueue()
|
| D | nccl_manager.cc | 170 reduction_op(reduction_op_in), in Collective()
191 const ncclRedOp_t reduction_op; // applies when <type> is a reduction. member
434 ncclRedOp_t reduction_op) { in AddToAllReduce() argument
435 AddParticipant(std::move(participant), context, kAllReduce, reduction_op); in AddToAllReduce()
459 ncclRedOp_t reduction_op) { in AddReduceSend() argument
460 AddParticipant(std::move(participant), context, kReduce, reduction_op); in AddReduceSend()
465 ncclRedOp_t reduction_op) { in AddReduceRecv() argument
467 AddParticipant(std::move(participant), context, kReduce, reduction_op); in AddReduceRecv()
492 ncclRedOp_t reduction_op) { in AddParticipant() argument
509 context.collective_key, data_type, collective_type, reduction_op, in AddParticipant()
[all …]
|
| D | nccl_manager_test.cc | 105 ncclRedOp_t reduction_op, TensorShape shape, in MakeReductionTestCase() argument
109 if (reduction_op == ncclProd) { in MakeReductionTestCase()
112 } else if (reduction_op == ncclSum) { in MakeReductionTestCase()
115 } else if (reduction_op == ncclMax) { in MakeReductionTestCase()
117 } else if (reduction_op == ncclMin) { in MakeReductionTestCase()
120 LOG(FATAL) << "Invalid reduction_op " << reduction_op; in MakeReductionTestCase()
136 if (reduction_op == ncclProd) { in MakeReductionTestCase()
138 } else if (reduction_op == ncclSum) { in MakeReductionTestCase()
140 } else if (reduction_op == ncclMax) { in MakeReductionTestCase()
144 } else if (reduction_op == ncclMin) { in MakeReductionTestCase()
[all …]
|
| D | nccl_manager.h | 166 const Context& context, ncclRedOp_t reduction_op);
182 const Context& context, ncclRedOp_t reduction_op);
184 const Context& context, ncclRedOp_t reduction_op);
228 ncclRedOp_t reduction_op);
|
| /external/mesa3d/src/compiler/spirv/ |
| D | vtn_subgroup.c | 405 nir_op reduction_op; in vtn_handle_subgroup() local
410 reduction_op = nir_op_iadd; in vtn_handle_subgroup()
415 reduction_op = nir_op_fadd; in vtn_handle_subgroup()
418 reduction_op = nir_op_imul; in vtn_handle_subgroup()
421 reduction_op = nir_op_fmul; in vtn_handle_subgroup()
426 reduction_op = nir_op_imin; in vtn_handle_subgroup()
431 reduction_op = nir_op_umin; in vtn_handle_subgroup()
436 reduction_op = nir_op_fmin; in vtn_handle_subgroup()
441 reduction_op = nir_op_imax; in vtn_handle_subgroup()
446 reduction_op = nir_op_umax; in vtn_handle_subgroup()
[all …]
|
| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | nccl_all_reduce_thunk.cc | 45 mlir::Operation* reduction_op = &block.front(); in MatchReductionComputation() local
46 if (reduction_op->getNumOperands() != 2 || reduction_op->getNumResults() != 1) in MatchReductionComputation()
49 reduction_op->getOperand(0).dyn_cast<mlir::BlockArgument>(); in MatchReductionComputation()
51 reduction_op->getOperand(1).dyn_cast<mlir::BlockArgument>(); in MatchReductionComputation()
52 mlir::OpResult result = reduction_op->getResult(0); in MatchReductionComputation()
59 StatusOr<HloOpcode> opcode = MhloToHloOpcode(reduction_op); in MatchReductionComputation()
|
| /external/tensorflow/tensorflow/core/kernels/ |
| D | segment_reduction_ops_test.cc | 67 std::unique_ptr<OpKernel> reduction_op( in BM_SegmentReduction() local
74 params.op_kernel = reduction_op.get(); in BM_SegmentReduction()
81 reduction_op->Compute(reduction_context.get()); in BM_SegmentReduction()
85 reduction_op->Compute(reduction_context.get()); in BM_SegmentReduction()
|
| D | nccl_ops.cc | 84 ncclRedOp_t reduction_op() const { return reduction_op_; } in reduction_op() function in tensorflow::__anonc894cd950111::NcclReduceOpBase
119 reduction_op()); in ComputeAsync()
150 reduction_op()); in ComputeAsync()
186 reduction_op()); in ComputeAsync()
|
| D | dynamic_partition_op_gpu.cu.cc | 416 gpuprim::Sum reduction_op; in CountAndSortParts() local
429 aggregates_out_it, num_runs_ptr, reduction_op, N, cu_stream); in CountAndSortParts()
443 unique_out_it, values_in, aggregates_out_it, num_runs_ptr, reduction_op, in CountAndSortParts()
|
| /external/tensorflow/tensorflow/python/ops/parallel_for/ |
| D | math_test.py | 491 def test_unsorted_segment_reduction(self, reduction_op): argument
505 return (reduction_op(data, seg_ids, num_segments),
506 reduction_op(data_0, seg_ids, num_segments),
507 reduction_op(data, seg_ids_0, num_segments))
|
| D | pfor.py | 1177 reduction_op = array_ops.identity_n(pl_outputs)[0].op
1178 self._reduce_map[reduction_op] = (concrete_function, args)
1179 if len(reduction_op.outputs) == 1:
1180 return reduction_op.outputs[0]
1182 return tuple(reduction_op.outputs)
|
| /external/mesa3d/src/compiler/nir/ |
| D | nir_print.c | 880 nir_op reduction_op = nir_intrinsic_reduction_op(instr); in print_intrinsic_instr() local
881 fprintf(fp, " reduction_op=%s", nir_op_infos[reduction_op].name); in print_intrinsic_instr()
|
| D | nir.h | 2072 INTRINSIC_IDX_ACCESSORS(reduction_op, REDUCTION_OP, unsigned) in INTRINSIC_IDX_ACCESSORS()
|
| /external/tensorflow/tensorflow/compiler/mlir/tensorflow/ir/ |
| D | tf_ops_a_m.cc | 2777 auto reduction_op = in FoldOperandsPermutation() local
2779 if (!reduction_op) return failure(); in FoldOperandsPermutation()
2781 auto reductions_value = reduction_op.value().dyn_cast<DenseElementsAttr>(); in FoldOperandsPermutation()
|