| /external/tensorflow/tensorflow/lite/delegates/gpu/common/task/ |
| D | work_group_picking.cc | 38 std::vector<int3> work_groups; in GenerateWorkGroupSizesXYMultipleOf() local
39 work_groups.reserve(32); in GenerateWorkGroupSizesXYMultipleOf()
57 work_groups.push_back({x, y, z}); in GenerateWorkGroupSizesXYMultipleOf()
62 return work_groups; in GenerateWorkGroupSizesXYMultipleOf()
68 std::vector<int3> work_groups; in GenerateWorkGroupSizesXMultipleOf() local
69 work_groups.reserve(32); in GenerateWorkGroupSizesXMultipleOf()
84 work_groups.push_back({x, y, z}); in GenerateWorkGroupSizesXMultipleOf()
89 return work_groups; in GenerateWorkGroupSizesXMultipleOf()
94 std::vector<int3>* work_groups) { in GetWorkGroupsAlignedToGrid() argument
100 grid, max_wg_size, kernel_info.max_work_group_size, work_groups); in GetWorkGroupsAlignedToGrid()
[all …]
|
| D | work_group_picking.h | 35 std::vector<int3>* work_groups);
41 std::vector<int3>* work_groups);
52 std::vector<int3>* work_groups);
56 std::vector<int3>* work_groups);
|
| D | gpu_operation.cc | 211 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
213 work_groups); in GetPossibleKernelWorkGroups()
|
| D | gpu_operation.h | 115 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const;
|
| /external/tensorflow/tensorflow/lite/delegates/gpu/common/ |
| D | workgroup_selection.cc | 36 std::vector<T>* work_groups) { in AddCornerCases() argument
60 work_groups->push_back({wg_x, wg_y, wg_z}); in AddCornerCases()
83 work_groups->push_back({x, y, z}); in AddCornerCases()
150 std::vector<T> work_groups; in GenerateWorkGroupSizes() local
151 work_groups.reserve(64); in GenerateWorkGroupSizes()
167 work_groups.push_back({x, y, z}); in GenerateWorkGroupSizes()
172 return work_groups; in GenerateWorkGroupSizes()
193 std::vector<T>* work_groups) { in GenerateWorkGroupSizesAlignedToGrid() argument
195 *work_groups = GenerateWorkGroupSizes<T>( in GenerateWorkGroupSizesAlignedToGrid()
199 if (work_groups->empty()) { in GenerateWorkGroupSizesAlignedToGrid()
[all …]
|
| D | workgroup_selection.h | 45 std::vector<T>* work_groups);
|
| /external/tensorflow/tensorflow/lite/delegates/gpu/common/tasks/ |
| D | winograd.h | 37 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
38 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
71 std::vector<int3>* work_groups) const override;
104 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
105 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
137 std::vector<int3>* work_groups) const override;
|
| D | mean_stddev_normalization.h | 37 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
38 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| D | softmax1x1.h | 32 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
33 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| D | winograd.cc | 469 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
471 work_groups->push_back(int3(4, 6, 1)); in GetPossibleKernelWorkGroups()
477 work_groups); in GetPossibleKernelWorkGroups()
481 work_groups->push_back(SelectBestWorkGroup(kernel_info)); in GetPossibleKernelWorkGroups()
683 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
685 work_groups->push_back(int3(8, 4, 1)); in GetPossibleKernelWorkGroups()
691 work_groups); in GetPossibleKernelWorkGroups()
695 work_groups->push_back(SelectBestWorkGroup(kernel_info)); in GetPossibleKernelWorkGroups()
|
| D | convolution_transposed_4x4.h | 43 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
44 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| D | conv_metal.h | 60 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
61 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| D | depthwise_conv_3x3_stride_h2.cc | 199 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
201 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
204 work_groups); in GetPossibleKernelWorkGroups()
|
| D | depthwise_conv_3x3.cc | 293 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
295 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
298 work_groups); in GetPossibleKernelWorkGroups()
|
| D | fully_connected.h | 122 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
123 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| D | reduce.h | 39 std::vector<int3>* work_groups) const override;
|
| D | convolution_transposed_3x3.cc | 354 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
357 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
361 work_groups); in GetPossibleKernelWorkGroups()
|
| D | convolution_transposed_3x3.h | 43 std::vector<int3>* work_groups) const override;
|
| D | reduce.cc | 466 std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
468 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
471 work_groups); in GetPossibleKernelWorkGroups()
|
| D | depthwise_conv_3x3_stride_h2.h | 41 std::vector<int3>* work_groups) const override;
|
| D | depthwise_conv_3x3.h | 42 std::vector<int3>* work_groups) const override;
|
| D | conv_buffer_1x1.cc | 316 const KernelInfo& kernel_info, std::vector<int3>* work_groups) const { in GetPossibleKernelWorkGroups()
318 work_groups); in GetPossibleKernelWorkGroups()
|
| D | conv_buffer_1x1.h | 48 std::vector<int3>* work_groups) const override;
|
| /external/tensorflow/tensorflow/lite/delegates/gpu/common/tasks/special/ |
| D | fc_fc_add.h | 98 std::vector<int3>* work_groups) const override { in GetPossibleKernelWorkGroups() argument
99 work_groups->push_back(work_group_size_); in GetPossibleKernelWorkGroups()
|
| /external/tensorflow/tensorflow/lite/delegates/gpu/cl/kernels/ |
| D | converter.cc | 56 std::vector<int3> work_groups; in DispatchKernel() local
58 &work_groups); in DispatchKernel()
59 const int3 work_group_size = work_groups[0]; in DispatchKernel()
|