| /third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp16/ |
| D | convolution_winograd_fp16.cc | 91 …thread_count_ * row_tile_ * input_unit_ * input_unit_ * conv_param_->input_channel_ * sizeof(float… in InitTmpBuffer()
99 …thread_count_ * row_tile_ * input_unit_ * input_unit_ * UP_ROUND(channel_out, C8NUM) * sizeof(floa… in InitTmpBuffer()
106 ctx_->allocator->Malloc(thread_count_ * C8NUM * input_unit_ * input_unit_ * sizeof(float16_t))); in InitTmpBuffer()
113 …ctx_->allocator->Malloc(thread_count_ * row_tile_ * conv_param_->input_channel_ * sizeof(float16_t… in InitTmpBuffer()
175 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(cal_plane, C8NUM)); in AdjustNumberOfThread()
176 if (thread_count_ <= 0) { in AdjustNumberOfThread()
180 conv_param_->thread_num_ = thread_count_; in AdjustNumberOfThread()
245 ret = ParallelLaunch(this->ms_context_, ConvolutionWinogradFp16Impl, this, thread_count_); in Run()
|
| D | convolution_1x1_fp16.cc | 55 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->row_, row_tile_)); in InitConv1x1Param()
56 if (thread_count_ <= 0) { in InitConv1x1Param()
60 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->row_, row_tile_), thread_count_) * row_tile_; in InitConv1x1Param()
63 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->col_, col_tile_)); in InitConv1x1Param()
64 if (thread_count_ <= 0) { in InitConv1x1Param()
68 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->col_, col_tile_), thread_count_) * col_tile_; in InitConv1x1Param()
70 matmul_param_->op_parameter_.thread_num_ = thread_count_; in InitConv1x1Param()
309 ret = ParallelLaunch(this->ms_context_, Convolution1x1Fp16RunHw, this, thread_count_); in Run()
320 ret = ParallelLaunch(this->ms_context_, Convolution1x1Fp16RunOc, this, thread_count_); in Run()
|
| D | convolution_fp16.cc | 79 …ram_->kernel_h_ * conv_param_->kernel_w_ * conv_param_->input_channel_ * row_tile_ * thread_count_; in InitTmpBuffer()
127 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(out_plane, row_tile_)); in AdjustNumberOfThread()
128 conv_param_->thread_num_ = thread_count_; in AdjustNumberOfThread()
184 ret = ParallelLaunch(this->ms_context_, ConvolutionFp16Impl, this, thread_count_); in Run()
|
| /third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp32/ |
| D | convolution_1x1_fp32.cc | 68 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->row_, row_tile_)); in InitConv1x1Param()
69 if (thread_count_ <= 0) { in InitConv1x1Param()
73 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->row_, row_tile_), thread_count_) * row_tile_; in InitConv1x1Param()
76 thread_count_ = MSMIN(op_parameter_->thread_num_, UP_DIV(matmul_param_->col_, col_tile_)); in InitConv1x1Param()
77 if (thread_count_ <= 0) { in InitConv1x1Param()
81 thread_stride_ = UP_DIV(UP_DIV(matmul_param_->col_, col_tile_), thread_count_) * col_tile_; in InitConv1x1Param()
229 int pack_input_size = multi_thread_by_hw_ ? (thread_count_ * row_tile_ * matmul_param_->deep_) in Run()
251 ret = ParallelLaunch(this->ms_context_, Convolution1x1RunHw, this, thread_count_); in Run()
254 ret = ParallelLaunch(this->ms_context_, Convolution1x1Run, this, thread_count_); in Run()
|
| D | sparse_to_dense_fp32.h | 33 : InnerKernel(parameter, inputs, outputs, ctx), ctx_(ctx), thread_count_(ctx->thread_num_) { in SparseToDenseCPUKernel()
35 s2d_param->thread_num_ = thread_count_; in SparseToDenseCPUKernel()
48 int thread_count_; variable
|
| D | local_response_norm_fp32.cc | 53 MS_CHECK_TRUE_RET(thread_count_ != 0, RET_ERROR); in DoLocalResponseNorm()
54 int stride = UP_DIV(outer_size, thread_count_); in DoLocalResponseNorm()
81 int error_code = ParallelLaunch(this->ms_context_, LocalResponseNormRun, this, thread_count_); in Run()
|
| D | convolution_winograd_fp32.cc | 43 …thread_count_ * tile_num_ * input_unit_ * input_unit_ * conv_param_->input_channel_ * sizeof(float… in InitTmpBuffer()
52 …ctx_->allocator->Malloc(thread_count_ * tile_num_ * input_unit_ * input_unit_ * oc8 * sizeof(float… in InitTmpBuffer()
59 ctx_->allocator->Malloc(thread_count_ * C4NUM * input_unit_ * input_unit_ * sizeof(float))); in InitTmpBuffer()
66 … ctx_->allocator->Malloc(thread_count_ * tile_num_ * conv_param_->input_channel_ * sizeof(float))); in InitTmpBuffer()
179 ret = ParallelLaunch(this->ms_context_, ConvolutionWinogradImpl, this, thread_count_); in Run()
|
| D | local_response_norm_fp32.h | 29 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in LocalResponseNormCPUKernel()
38 int thread_count_;
|
| /third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/int8/ |
| D | unsqueeze_int8.h | 32 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) { in Unsqueezeint8CPUKernel()
34 param_->thread_count_ = op_parameter_->thread_num_; in Unsqueezeint8CPUKernel()
50 int thread_count_{0};
|
| D | mul_int8.cc | 157 count_unit_ = thread_count_ > 1 ? UP_DIV(elements_num_, thread_count_) : elements_num_; in Run()
158 return ParallelLaunch(this->ms_context_, FastHWBroadcastMulInt8Run, this, thread_count_); in Run()
162 count_unit_ = thread_count_ > 1 ? UP_DIV(elements_num_, thread_count_) : elements_num_; in Run()
178 ret = ParallelLaunch(this->ms_context_, MulInt8Run, this, thread_count_); in Run()
184 ret = ParallelLaunch(this->ms_context_, MulInt8Run, this, thread_count_); in Run()
|
| D | tanh_int8.cc | 40 thread_count_ = MSMIN(element_size_, op_parameter_->thread_num_); in ReSize()
41 thread_stride_ = UP_DIV(element_size_, thread_count_); in ReSize()
73 auto ret = ParallelLaunch(this->ms_context_, TanhInt8Run, this, thread_count_); in Run()
|
| D | add_int8.cc | 97 thread_count_ = MSMIN(elements_num_, op_parameter_->thread_num_); in ReSize()
166 if (thread_count_ == 0) { in BroadcastRun()
170 int stride = UP_DIV(out_size_, thread_count_); in BroadcastRun()
204 if (thread_count_ == 0) { in DoExecute()
208 int stride = UP_DIV(elements_num_, thread_count_); in DoExecute()
243 auto ret = ParallelLaunch(this->ms_context_, AddInt8Run, this, thread_count_); in Run()
|
| D | convolution_int8.cc | 132 input_sum_size = static_cast<size_t>(up_round_oc * tile_num_ * thread_count_) * sizeof(int32_t); in InitWeightBias()
134 input_sum_size = static_cast<size_t>(tile_num_ * thread_count_) * sizeof(int32_t); in InitWeightBias()
155 … ctx_->allocator->Malloc(thread_count_ * tile_num_ * kernel_plane * conv_param_->input_channel_)); in InitTmpBuffer()
160 …packed_input_ = reinterpret_cast<int8_t *>(ctx_->allocator->Malloc(tmp_size * thread_count_ * tile… in InitTmpBuffer()
230 int error_code = ParallelLaunch(this->ms_context_, ConvolutionInt8Impl, this, thread_count_); in Run()
|
| D | unsqueeze_int8.cc | 43 param_->thread_count_ = thread_count_; in Init()
52 thread_sz_count_ = MSMIN(thread_count_, data_size_); in ReSize()
|
| /third_party/boost/doc/html/boost_asio/example/cpp11/executors/ |
| D | fork_join.cpp | 33 for (thread_count_ = 0; thread_count_ < thread_count; ++thread_count_) in fork_join_pool()
126 if (queue_.size() > thread_count_ * 16) in do_execute()
152 std::size_t thread_count_; member in fork_join_pool
|
| /third_party/boost/libs/asio/example/cpp14/executors/ |
| D | fork_join.cpp | 33 for (thread_count_ = 0; thread_count_ < thread_count; ++thread_count_) in fork_join_pool()
126 if (queue_.size() > thread_count_ * 16) in do_execute()
152 std::size_t thread_count_; member in fork_join_pool
|
| /third_party/boost/libs/asio/example/cpp11/executors/ |
| D | fork_join.cpp | 33 for (thread_count_ = 0; thread_count_ < thread_count; ++thread_count_) in fork_join_pool()
126 if (queue_.size() > thread_count_ * 16) in do_execute()
152 std::size_t thread_count_; member in fork_join_pool
|
| /third_party/mindspore/mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/fp16/ |
| D | crop_fp16.c | 45 const int thread_count = para->thread_count_; in Fp16Crop1D()
64 const int thread_count = para->thread_count_; in Fp16Crop2D()
90 const int thread_count = para->thread_count_; in Fp16Crop3D()
126 const int thread_count = para->thread_count_; in Fp16Crop4D()
|
| /third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp32_grad/ |
| D | assign.h | 29 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in AssignCPUKernel()
37 int thread_count_ = 1;
|
| D | neg_grad.h | 30 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in NegGradCPUKernel()
38 int thread_count_;
|
| D | dropout.h | 27 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in DropoutCPUKernel()
38 int thread_count_ = 1; variable
|
| D | dropout_grad.h | 27 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in DropoutGradCPUKernel()
38 int thread_count_ = 1; variable
|
| D | arithmetic_self_grad.h | 32 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in ArithmeticSelfGradCPUKernel()
40 int thread_count_;
|
| /third_party/mindspore/mindspore/lite/src/runtime/kernel/arm/fp16_grad/ |
| D | neg_fp16_grad.h | 30 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in NegGradCPUKernelFp16()
38 int thread_count_;
|
| D | dropout_fp16_grad.h | 27 : InnerKernel(parameter, inputs, outputs, ctx), thread_count_(ctx->thread_num_) {} in DropoutGradCPUKernelFp16()
38 int thread_count_ = 1; variable
|