Home
last modified time | relevance | path

Searched full:kernel (Results 1 – 25 of 8191) sorted by relevance

12345678910>>...328

/third_party/mindspore/mindspore-src/source/.jenkins/check/config/
Dwhitelizard.txt32 mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/math/unary_op_gpu_kernel.h:mindspore::kernel::Un…
34 mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/drop_out_gen_mask_kernels.cc:…
72 mindspore/mindspore/lite/src/litert/kernel/opencl/kernel/conv2d.cc:mindspore::kernel::UseWinograd4x…
75 mindspore/mindspore/lite/src/litert/ios_reg_kernels.h:mindspore::kernel::IosRegisterKernels
76 mindspore/mindspore/lite/src/litert/kernel/cpu/base/quant_dtype_cast.cc:mindspore::kernel::QuantDTy…
77 mindspore/mindspore/lite/src/litert/kernel/cpu/base/quant_dtype_cast.cc:mindspore::kernel::QuantDTy…
78 mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/infer/strided_slice_infer.c:StridedSliceIn…
79 mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/infer/lstm_infer.c:CheckInputShapeValid
80 mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/fp16/winograd_transform_fp16.c:WinogradInp…
81 mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/fp16/pooling_fp16.c:AvgPoolingFp16
[all …]
Dfilter_cppcheck.txt10 "mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/drop_out_gen_mask_kernels.cc…
18 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sequence/sequence_concat_cpu_kernel.cc" …
19 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sparse_apply_ftrl_cpu_kernel.cc" …
20 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sparse_apply_adam_cpu_kernel.cc" …
21 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sparse_apply_lazy_adam_cpu_kernel.cc" …
22 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sparse_apply_proximal_adagrad_cpu_kernel.cc" …
24 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/embedding_look_up_cpu_kernel.cc" …
28 "mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/arrays/cast_gpu_kernel.cc" …
66 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/arithmetic_cpu_kernel.cc" …
68 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/xlogy_cpu_kernel.cc" …
[all …]
Dfilter_cpplint.txt12 "mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/arrays/broadcast_to_gpu_kernel.cc" …
17 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/" …
22 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/arithmetic_self_cpu_kernel.cc" …
24 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/generator_cpu_kernel.cc" "buil…
29 "mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/gather_grad_kernels.cc" …
30 "mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/drop_out_gen_mask_kernels.cc…
32 "mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/cuda_impl/cuda_ops/concatv2_impl.cu" …
36 "mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/sequence/sequence_addn_gpu_kernel.cc" …
37 "mindspore/mindspore/ccsrc/plugin/device/cpu/kernel/sequence/sequence_addn_cpu_kernel.cc" …
39 "mindspore/mindspore/ccsrc/plugin/device/gpu/kernel/math/binary_ops_gpu_kernel.cc" …
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/ccsrc/plugin/device/gpu/hal/device/
Dgpu_kernel_build.cc20 #include "kernel/kernel.h"
24 #include "plugin/device/gpu/kernel/akg/akg_gpu_kernel_build.h"
30 #include "kernel/framework_utils.h"
32 #include "plugin/device/gpu/kernel/gpu_kernel_factory.h"
37 void SetGpuRefMapToKernelInfo(const CNodePtr &apply_kernel, const std::vector<kernel::KernelAttr> &… in SetGpuRefMapToKernelInfo()
43 auto kernel_attr = kernel::GetKernelAttrFromNode(apply_kernel); in SetGpuRefMapToKernelInfo()
44 auto [is_match, index] = kernel::MatchKernelAttr(kernel_attr, kernel_attrs); in SetGpuRefMapToKernelInfo()
51 << " does not support this kernel data type: " << kernel_attr; in SetGpuRefMapToKernelInfo()
64 kernel::KernelMeta *bin_map = kernel::KernelMeta::GetInstance(); in CreateGPUKernel()
68 for (auto kernel : kernels) { in CreateGPUKernel() local
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/control_flow/
Dcontrol_flow_scheduler.h42 int Schedule(std::vector<kernel::KernelExec *> *dst_kernels);
43 …void SetSubgraphForPartialNode(std::unordered_map<kernel::KernelExec *, size_t> *partial_kernel_su…
44 … std::unordered_map<size_t, kernel::KernelExec *> *subgraph_index_subgraph_kernel_map);
45 std::vector<kernel::KernelExec *> GetNonTailCalls() const { return non_tail_calls_; } in GetNonTailCalls()
46 void RecordSubgraphCaller(const size_t &subgraph_index, kernel::KernelExec *partial_node);
49 int SplitNonTailCallSubGraphs(std::vector<kernel::KernelExec *> *dst_kernels);
50 …// We insert entrance subgraph kernel and exit subgraph kernel define the boundary of the subgraph.
51 int BuildBoundaryForMultipleCalledGraph(std::vector<kernel::KernelExec *> *dst_kernels);
53 int IsolateOutputForCallOutputGraph(std::vector<kernel::KernelExec *> *dst_kernels);
57 int IsolateSameInputPartials(std::vector<kernel::KernelExec *> *dst_kernels);
[all …]
Dcontrol_actor_creator.cc26 std::shared_ptr<LiteOpActor> CreateActor(kernel::KernelExec *kernel, lite::InnerContext *ctx) { in CreateActor() argument
28 if (kernel::KernelExecUtil::IsSwitchTypeCall(kernel)) { in CreateActor()
29 actor = std::make_shared<LiteSwitchOpActor>(kernel, ctx); in CreateActor()
30 } else if (kernel->subgraph_type() == kernel::kEntranceSubGraph) { in CreateActor()
31 actor = std::make_shared<LiteEntranceOpActor>(kernel, ctx); in CreateActor()
32 } else if (kernel->subgraph_type() == kernel::kExitSubGraph) { in CreateActor()
33 actor = std::make_shared<LiteExitOpActor>(kernel, ctx); in CreateActor()
34 } else if (kernel->subgraph_type() != kernel::kNotSubGraph) { in CreateActor()
35 auto subgraph_kernel = reinterpret_cast<kernel::SubGraphKernel *>(kernel); in CreateActor()
37 …(kernel->subgraph_type() == kernel::kCpuFP32SubGraph || kernel->subgraph_type() == kernel::kCpuFP1… in CreateActor()
[all …]
Dcontrol_flow_scheduler.cc22 #include "src/litert/kernel/cpu/base/partial_fusion.h"
24 #include "src/control_flow/kernel/exit_subgraph_kernel.h"
25 #include "src/control_flow/kernel/identity_kernel.h"
36 int ControlFlowScheduler::Schedule(std::vector<kernel::KernelExec *> *dst_kernels) { in Schedule()
52 int ControlFlowScheduler::SplitNonTailCallSubGraphs(std::vector<kernel::KernelExec *> *dst_kernels)… in SplitNonTailCallSubGraphs()
53 std::set<kernel::KernelExec *> all_non_tail_subgraphs = GetNonTailCallSubGraphs(dst_kernels); in SplitNonTailCallSubGraphs()
61 auto subgraph_kernel = reinterpret_cast<kernel::SubGraphKernel *>(cur); in SplitNonTailCallSubGraphs()
63 MS_LOG(ERROR) << "kernel is not a subgraph kernel"; in SplitNonTailCallSubGraphs()
66 std::vector<kernel::KernelExec *> new_subgraphs{}; in SplitNonTailCallSubGraphs()
77 auto partial_kernel = reinterpret_cast<kernel::PartialFusionKernel *>(partial_node->kernel()); in SplitNonTailCallSubGraphs()
[all …]
/third_party/ltp/tools/
Drestore_kernel_faults_default.sh35 echo 0 > /sys/kernel/debug/fail_io_timeout/reject-end
36 echo 0 > /sys/kernel/debug/fail_io_timeout/reject-start
37 echo 4294967295 > /sys/kernel/debug/fail_io_timeout/require-end
38 echo 0 > /sys/kernel/debug/fail_io_timeout/require-start
39 echo 32 > /sys/kernel/debug/fail_io_timeout/stacktrace-depth
40 echo N > /sys/kernel/debug/fail_io_timeout/task-filter
41 echo 2 > /sys/kernel/debug/fail_io_timeout/verbose
42 echo 0 > /sys/kernel/debug/fail_io_timeout/space
43 echo 1 > /sys/kernel/debug/fail_io_timeout/times
44 echo 1 > /sys/kernel/debug/fail_io_timeout/interval
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/ccsrc/plugin/device/cpu/kernel/nnacl/kernel/
Dconvolution_winograd.c17 #include "nnacl/kernel/convolution_winograd.h"
18 #include "nnacl/kernel/convolution_winograd_base.h"
20 #include "nnacl/kernel/convolution_winograd_avx.h"
23 #include "nnacl/kernel/convolution_winograd_sse.h"
26 #include "nnacl/kernel/convolution_winograd_arm64.h"
29 #include "nnacl/kernel/convolution_winograd_arm32.h"
33 ConvolutionWinogradBaseStruct *kernel = NULL; in SelectConvolutionWinograd() local
36 kernel = CreateConvWinogradAVX(conv_param); in SelectConvolutionWinograd()
37 if (kernel != NULL) { in SelectConvolutionWinograd()
38 return kernel; in SelectConvolutionWinograd()
[all …]
Dconvolution_im2col.c17 #include "nnacl/kernel/convolution_im2col.h"
18 #include "nnacl/kernel/convolution_im2col_base.h"
20 #include "nnacl/kernel/convolution_im2col_arm32.h"
23 #include "nnacl/kernel/convolution_im2col_arm64.h"
26 #include "nnacl/kernel/convolution_im2col_sse.h"
29 #include "nnacl/kernel/convolution_im2col_avx.h"
33 #include "nnacl/kernel/convolution_im2col_avx512.h"
37 ConvolutionBaseStruct *kernel = NULL; in CreateConvolutionIm2Col() local
43 kernel = CreateConvIm2ColAVX512(conv_param); in CreateConvolutionIm2Col()
44 if (kernel != NULL) { in CreateConvolutionIm2Col()
[all …]
/third_party/ltp/
DREADME.kernel_config2 Enable CODE COVERAGE for your Kernel:
4 1) Apply kernel-gcov patch(s) against the corresponding Kernel. They are available here:
5 http://ltp.cvs.sourceforge.net/viewvc/ltp/utils/analysis/gcov-kernel/linux-2.*.*-gcov.patch?view=lo…
6 2) Also enable the following options in your .config file before building the kernel
13 Enabling Kernel Configuration to test Containers/Namespaces
32 Enabling Kernel Configuration to test Controllers
44 Enabling Kernel Configuration to test Power Management features
66 Enabling Kernel Configuration to test filecaps security feature
70 Enabling Kernel Configuration to test SELinux security feature
72 Your Kernel should have been built with the following options to
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/kernel/cpu/int8/
Dconvolution_int8_creator.cc17 #include "src/litert/kernel/cpu/int8/convolution_int8_creator.h"
18 #include "src/litert/kernel/cpu/int8/convolution_int8.h"
19 #include "src/litert/kernel/cpu/int8/convolution_1x1_int8.h"
20 #include "src/litert/kernel/cpu/int8/convolution_3x3_int8.h"
21 #include "src/litert/kernel/cpu/int8/convolution_depthwise_int8.h"
22 #include "src/litert/kernel/cpu/int8/convolution_depthwise_3x3_int8.h"
23 #include "src/litert/kernel/cpu/int8/convolution_depthwise_slidewindow_int8.h"
24 #include "src/litert/kernel/cpu/int8/group_convolution_int8.h"
25 #include "src/litert/kernel/cpu/base/group_convolution_creator.h"
30 using mindspore::kernel::KERNEL_ARCH;
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/kernel/cpu/fp16/
Dconvolution_delegate_fp16.cc17 #include "src/litert/kernel/cpu/fp16/convolution_delegate_fp16.h"
19 #include "src/litert/kernel/cpu/fp16/convolution_fp16.h"
20 #include "src/litert/kernel/cpu/fp16/convolution_winograd_fp16.h"
21 #include "src/litert/kernel/cpu/fp16/convolution_1x1_fp16.h"
22 #include "src/litert/kernel/cpu/fp16/group_convolution_fp16.h"
23 #include "src/litert/kernel/cpu/fp16/convolution_depthwise_fp16.h"
24 #include "src/litert/kernel/cpu/fp16/convolution_depthwise_slidewindow_fp16.h"
25 #include "src/litert/kernel/cpu/fp16/convolution_depthwise_3x3_fp16.h"
26 #include "src/litert/kernel/cpu/base/group_convolution_creator.h"
32 using mindspore::kernel::KERNEL_ARCH;
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/kernel/cpu/fp32/
Dconvolution_delegate_fp32.cc17 #include "src/litert/kernel/cpu/fp32/convolution_delegate_fp32.h"
19 #include "src/litert/kernel/cpu/fp32/convolution_im2col_fp32.h"
20 #include "src/litert/kernel/cpu/fp32/convolution_1x1_fp32.h"
21 #include "src/litert/kernel/cpu/fp32/convolution_winograd_fp32.h"
22 #include "src/litert/kernel/cpu/fp32/convolution_depthwise_fp32.h"
23 #include "src/litert/kernel/cpu/fp32/convolution_depthwise_slidewindow_fp32.h"
24 #include "src/litert/kernel/cpu/fp32/convolution_depthwise_slidewindow_x86_fp32.h"
25 #include "src/litert/kernel/cpu/base/group_convolution_creator.h"
26 #include "src/litert/kernel/cpu/fp32/group_convolution_fp32.h"
27 #include "src/litert/kernel/cpu/fp32/convolution_sw_1x1_fp32.h"
[all …]
Dconvolution_im2col_fp32.cc17 #include "src/litert/kernel/cpu/fp32/convolution_im2col_fp32.h"
18 #include "src/litert/kernel/cpu/fp32/convolution_im2col_base_fp32.h"
20 #include "src/litert/kernel/cpu/fp32/convolution_im2col_avx512_fp32.h"
23 #include "src/litert/kernel/cpu/fp32/convolution_im2col_avx_fp32.h"
27 #include "src/litert/kernel/cpu/fp32/convolution_im2col_sse_fp32.h"
31 #include "src/litert/kernel/cpu/fp32/convolution_im2col_arm32_fp32.h"
35 #include "src/litert/kernel/cpu/fp32/convolution_im2col_arm64_fp32.h"
39 namespace mindspore::kernel { namespace
43 LiteKernel *kernel = nullptr; in CreateConvolutionIm2ColCPUKernel() local
45 if (kernel == nullptr && outputs.front()->format() != NC4HW4) { in CreateConvolutionIm2ColCPUKernel()
[all …]
/third_party/mindspore/mindspore-src/source/include/api/
Ddelegate.h26 #include "include/api/kernel.h"
36 using KernelIter = std::vector<kernel::Kernel *>::iterator;
43 DelegateModel(std::vector<kernel::Kernel *> *kernels, const std::vector<MSTensor> &inputs, in DelegateModel()
44 … const std::vector<MSTensor> &outputs, const std::map<kernel::Kernel *, const T *> &primitives, in DelegateModel() argument
51 /// \brief Get Primitive of kernel::Kernel.
53 /// \param[in] kernel kernel in DelegateModel kernels vector.
55 /// \return The Primitive of The kernel.
56 const T *GetPrimitive(kernel::Kernel *kernel) const { in GetPrimitive() argument
57 if (primitives_.find(kernel) != primitives_.end()) { in GetPrimitive()
58 return primitives_.at(kernel); in GetPrimitive()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/ccsrc/plugin/device/cpu/kernel/
Dbatch_norm_grad_grad_cpu_kernel.h25 #include "plugin/device/cpu/kernel/cpu_kernel.h"
29 namespace kernel {
49 void TrainingComputeNHWC(const std::vector<kernel::KernelTensor *> &inputs,
50 const std::vector<kernel::KernelTensor *> &workspace,
51 const std::vector<kernel::KernelTensor *> &outputs) const;
54 void InferenceComputeNHWC(const std::vector<kernel::KernelTensor *> &inputs,
55 const std::vector<kernel::KernelTensor *> &workspace,
56 const std::vector<kernel::KernelTensor *> &outputs) const;
59 void TrainingComputeNCHW(const std::vector<kernel::KernelTensor *> &inputs,
60 const std::vector<kernel::KernelTensor *> &workspace,
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/include/registry/
Dopencl_runtime_wrapper.h48 /// \param[in] kernel Used to return the compiled kernel
50 /// \param[in] kernel_name Define OpenCl source kernel name.
51 /// \param[in] build_options_ext Define OpenCl kernel build options.
53 /// \return Status as a status identification of build Kernel
54 …inline Status BuildKernel(cl::Kernel *kernel, const std::string &program_name, const std::string &…
57 /// \brief Set kernel argument
59 /// \param[in] kernel Define OpenCl kernel.
60 /// \param[in] index Define OpenCl kernel argument index.
61 /// \param[in] value Define OpenCl kernel argument value pointer.
62 /// \param[in] mem_type Define OpenCl kernel argument value memory type.
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/train/
Dtrain_session.cc53 void AddNonConstTrainableParams(const std::vector<kernel::KernelExec *> &in_kernels, kernel::Optimi… in AddNonConstTrainableParams()
128 for (auto kernel : this->train_kernels_) { in AllocWorkSpace()
129 if (workspace_size < static_cast<kernel::LiteKernel *>(kernel->kernel())->workspace_size()) { in AllocWorkSpace()
130 workspace_size = static_cast<kernel::LiteKernel *>(kernel->kernel())->workspace_size(); in AllocWorkSpace()
138 for (auto kernel : this->train_kernels_) { in AllocWorkSpace()
139 static_cast<kernel::LiteKernel *>(kernel->kernel())->set_workspace(workspace_); in AllocWorkSpace()
149 for (auto kernel : this->train_kernels_) { in FreeWorkSpace()
150 static_cast<kernel::LiteKernel *>(kernel->kernel())->FreeWorkspace(); in FreeWorkSpace()
207 int TrainSession::AllocTensors(const std::vector<kernel::KernelExec *> &kernels) { in AllocTensors()
214 for (auto &kernel : kernels) { in AllocTensors() local
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/pass/format_pass/
Deliminate_transpose.cc23 int TransFullyFusion(kernel::SubGraphKernel *subgraph, kernel::KernelExec *trans_kernel0, in TransFullyFusion()
24 kernel::KernelExec *trans_kernel1) { in TransFullyFusion()
30 auto in_kernel = kernel::KernelExecUtil::FindInKernelForInTensor(trans_kernel0, in_tensor); in TransFullyFusion()
31 auto out_kernels = kernel::KernelExecUtil::FindOutKernelsForOutTensor(trans_kernel1, out_tensor); in TransFullyFusion()
35 MS_LOG(ERROR) << "Update tensor failed when fusing kernel " << trans_kernel0->name() << " and " in TransFullyFusion()
47 int TransHeadTailFusion(kernel::SubGraphKernel *subgraph, kernel::KernelExec *trans_kernel0, in TransHeadTailFusion()
48 kernel::KernelExec *trans_kernel1, const TransInfoPair &trans_info, in TransHeadTailFusion()
57 auto in_kernel = kernel::KernelExecUtil::FindInKernelForInTensor(trans_kernel0, in_tensor); in TransHeadTailFusion()
58 auto out_kernels = kernel::KernelExecUtil::FindOutKernelsForOutTensor(trans_kernel1, out_tensor); in TransHeadTailFusion()
60 // new trans kernel: src_format -> dst_format in TransHeadTailFusion()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/extendrt/
Dexecution_flow.cc24 for (auto kernel : kernels_) { in ~ExecutionFlow() local
25 delete kernel; in ~ExecutionFlow()
30 kernel::SubGraphKernel *ExecutionFlow::ConstructFusionKernel() { in ConstructFusionKernel()
31 kernel::KernelExecUtil::FindAllInoutKernels(kernels_); in ConstructFusionKernel()
36 kernel::SubGraphType cur_sub_graph_type = this->GetSubGraphType(kernels_[0]); in ConstructFusionKernel()
41 …kernel::KernelExecUtil::CreateSubGraphKernel(kernels_, &inputs_, &outputs_, cur_sub_graph_type, *c… in ConstructFusionKernel()
65 for (const auto &kernel : kernels_) { in Dump() local
66 oss << kernel->ToString() << std::endl << std::endl; in Dump()
72 mindspore::kernel::SubGraphType ExecutionFlow::GetSubGraphType(abstract::Kernel *kernel) { in GetSubGraphType() argument
73 if (kernel == nullptr) { in GetSubGraphType()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/executor/
Dsub_graph_kernel.cc24 #include "src/litert/kernel/cpu/fp16/fp16_op_handler.h"
32 namespace mindspore::kernel { namespace
51 for (auto kernel : this->in_nodes_) { in ToString()
53 oss << kernel->ToString() << std::endl; in ToString()
56 for (auto kernel : this->out_nodes_) { in ToString()
58 oss << kernel->ToString() << std::endl; in ToString()
61 for (auto kernel : this->nodes_) { in ToString()
63 oss << kernel->ToString() << std::endl; in ToString()
84 for (auto kernel : nodes_) { in InferShape() local
85 MS_ASSERT(kernel != nullptr); in InferShape()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/
Dscheduler.cc51 #include "src/litert/kernel/cpu/fp16/fp16_op_handler.h"
54 #include "src/litert/kernel/opencl/opencl_subgraph.h"
55 #include "src/litert/kernel/gpu/opencl/opencl_runtime.h"
79 int CastKernelWeight(const kernel::SubGraphType &belong_subgraph_type, const kernel::KernelExec *ke… in CastKernelWeight() argument
81 MS_ASSERT(kernel != nullptr); in CastKernelWeight()
82 MS_ASSERT(kernel->subgraph_type() == kernel::kNotSubGraph); in CastKernelWeight()
83 …if (belong_subgraph_type != kernel::kCpuFP32SubGraph && belong_subgraph_type != kernel::kCpuFP16Su… in CastKernelWeight()
86 for (auto *tensor : kernel->in_tensors()) { in CastKernelWeight()
97 …if (tensor->data_type() == kNumberTypeFloat32 && belong_subgraph_type == kernel::kCpuFP16SubGraph)… in CastKernelWeight()
103 …} else if (tensor->data_type() == kNumberTypeFloat16 && belong_subgraph_type == kernel::kCpuFP32Su… in CastKernelWeight()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/lite/src/litert/kernel/gpu/opencl/
Dopencl_executor.cc17 #include "src/litert/kernel/gpu/opencl/opencl_executor.h"
18 #include "src/litert/kernel/opencl/utils.h"
24 … const std::vector<kernel::KernelExec *> &kernels, const KernelCallBack &before, in Run()
33 … const std::vector<kernel::KernelExec *> &kernels, const KernelCallBack &before, in RunOrTune()
41 for (auto *kernel : kernels) { in RunOrTune() local
42 MS_ASSERT(kernel); in RunOrTune()
44 callbackParam.node_name = kernel->name(); in RunOrTune()
45 callbackParam.node_type = kernel->type(); in RunOrTune()
46 … if ((before != nullptr) && !before(kernel->in_tensors(), kernel->out_tensors(), callbackParam)) { in RunOrTune()
47 MS_LOG(ERROR) << "run kernel before_callback failed, name: " << kernel->name(); in RunOrTune()
[all …]
/third_party/mindspore/mindspore-src/source/mindspore/ccsrc/backend/common/mem_reuse/
Dmem_swap_manager.cc31 for (const auto &kernel : execution_order_) { in Init() local
32 // Parse topo order of kernel in Init()
33 (void)kernel_execution_info_.emplace(kernel.get(), kernel_index++); in Init()
35 auto kernel_mod = AnfAlgo::GetKernelMod(kernel); in Init()
39 size_t output_num = AnfAlgo::GetOutputTensorNum(kernel); in Init()
41 TensorInfo tensor_info = {output_sizes[output_idx], kernel, output_idx}; in Init()
46 // Parse topo order of user kernel in Init()
129 const AnfNodePtr &kernel = tensor_info.kernel_; in CheckDistanceBetweenKernels() local
130 auto &kernel_exec_info = SearchKernelExecutionInfo(kernel); in CheckDistanceBetweenKernels()
153 const AnfNodePtr &kernel = tensor_info.kernel_; in CheckDistanceBetweenKernelsWithIdx() local
[all …]

12345678910>>...328