| /external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
| D | launch_dimensions.h | 30 class LaunchDimensions {
38 LaunchDimensions() in LaunchDimensions() function
41 LaunchDimensions(int64 block_x_count, int64 thread_x_count_per_block) in LaunchDimensions() function
45 LaunchDimensions(const Dim3D& block_counts, in LaunchDimensions() function
66 const LaunchDimensions& launch_dims);
68 LaunchDimensions CalculateLaunchDimensions(const Shape& shape,
|
| D | launch_dimensions.cc | 28 const LaunchDimensions& launch_dims) { in operator <<()
29 LaunchDimensions::Dim3D block_counts = launch_dims.block_counts(); in operator <<()
30 LaunchDimensions::Dim3D thread_counts = launch_dims.thread_counts_per_block(); in operator <<()
57 LaunchDimensions CalculateLaunchDimensions(const Shape& shape, in CalculateLaunchDimensions()
62 return LaunchDimensions(); in CalculateLaunchDimensions()
110 return LaunchDimensions(block_count, threads_per_block); in CalculateLaunchDimensions()
|
| D | parallel_loop_emitter.h | 36 const LaunchDimensions& launch_dimensions,
42 const LaunchDimensions& launch_dimensions,
52 const LaunchDimensions& launch_dimensions,
68 const LaunchDimensions launch_dimensions_;
|
| D | kernel_thunk.h | 58 void SetLaunchDimensions(const LaunchDimensions& launch_dims);
73 LaunchDimensions launch_dimensions_;
|
| D | parallel_loop_emitter.cc | 36 const LaunchDimensions& launch_dimensions, llvm::IRBuilder<>* b, in ParallelLoopEmitter()
45 const LaunchDimensions& launch_dimensions, llvm::IRBuilder<>* b, in ParallelLoopEmitter()
54 const LaunchDimensions& launch_dimensions, llvm::IRBuilder<>* b, in ParallelLoopEmitter()
|
| D | kernel_thunk.cc | 65 void KernelThunk::SetLaunchDimensions(const LaunchDimensions& launch_dims) { in SetLaunchDimensions()
90 LaunchDimensions launch_dimensions; in ExecuteOnStream()
|
| D | buffer_comparator.cc | 618 LaunchDimensions dim = in DeviceCompare()
621 LaunchDimensions::Dim3D thread_counts = dim.thread_counts_per_block(); in DeviceCompare()
622 LaunchDimensions::Dim3D block_counts = dim.block_counts(); in DeviceCompare()
|
| D | stream_executor_util.cc | 212 const LaunchDimensions& dims, se::Stream* stream) { in ExecuteKernelOnStream()
218 LaunchDimensions::Dim3D thread_counts = dims.thread_counts_per_block(); in ExecuteKernelOnStream()
219 LaunchDimensions::Dim3D block_counts = dims.block_counts(); in ExecuteKernelOnStream()
|
| D | stream_executor_util.h | 76 const LaunchDimensions& dims, se::Stream* stream);
|
| D | ir_emitter_unnested.cc | 158 void UpdateLaunchDimensions(const LaunchDimensions& launch_dims, Thunk* thunk, in UpdateLaunchDimensions()
924 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in EmitPadToStaticFromMlir()
1047 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in EmitSliceToDynamicFromMlir()
1841 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in EmitLoopFusionFromMlir()
1918 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in HandleFusion()
2042 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in HandleFusion()
2238 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in EmitSelectAndScatterFromMlir()
2722 LaunchDimensions launch_dimensions = CalculateLaunchDimensions( in EmitScatter()
2931 LaunchDimensions standard_launch_dimensions = CalculateLaunchDimensions( in EmitSortFromMlir()
2978 LaunchDimensions tiled_launch_dimensions(num_blocks, kThreadsPerBlock); in EmitSortFromMlir()
[all …]
|
| /external/llvm-project/parallel-libs/acxxel/ |
| D | opencl_acxxel.cpp | 123 KernelLaunchDimensions LaunchDimensions,
508 void *Stream, void *Kernel, KernelLaunchDimensions LaunchDimensions, in rawEnqueueKernelLaunch() argument
519 size_t LocalWorkSize[] = {LaunchDimensions.BlockX, LaunchDimensions.BlockY, in rawEnqueueKernelLaunch()
520 LaunchDimensions.BlockZ}; in rawEnqueueKernelLaunch()
521 size_t GlobalWorkSize[] = {LaunchDimensions.BlockX * LaunchDimensions.GridX, in rawEnqueueKernelLaunch()
522 LaunchDimensions.BlockY * LaunchDimensions.GridY, in rawEnqueueKernelLaunch()
523 LaunchDimensions.BlockZ * LaunchDimensions.GridZ}; in rawEnqueueKernelLaunch()
|
| D | cuda_acxxel.cpp | 140 KernelLaunchDimensions LaunchDimensions,
480 void *Stream, void *Kernel, KernelLaunchDimensions LaunchDimensions, in rawEnqueueKernelLaunch() argument
483 cuLaunchKernel(static_cast<CUfunction>(Kernel), LaunchDimensions.GridX, in rawEnqueueKernelLaunch()
484 LaunchDimensions.GridY, LaunchDimensions.GridZ, in rawEnqueueKernelLaunch()
485 LaunchDimensions.BlockX, LaunchDimensions.BlockY, in rawEnqueueKernelLaunch()
486 LaunchDimensions.BlockZ, SharedMemoryBytes, in rawEnqueueKernelLaunch()
|
| D | acxxel.cpp | 65 KernelLaunchDimensions LaunchDimensions, in asyncKernelLaunch() argument
70 TheHandle.get(), TheKernel.TheHandle.get(), LaunchDimensions, Arguments, in asyncKernelLaunch()
|
| D | acxxel.h | 359 KernelLaunchDimensions LaunchDimensions,
660 KernelLaunchDimensions LaunchDimensions,
|
| /external/tensorflow/tensorflow/compiler/xla/service/llvm_ir/ |
| D | dynamic_update_slice_util.cc | 98 const IrArray& output_array, const gpu::LaunchDimensions* launch_dimensions, in EmitDynamicUpdateSliceInPlaceImpl()
194 const gpu::LaunchDimensions* launch_dimensions, llvm::IRBuilder<>* b) { in EmitFusedDynamicUpdateSliceInPlaceImpl()
248 const gpu::LaunchDimensions& launch_dimensions, llvm::IRBuilder<>* b) { in EmitParallelFusedDynamicUpdateSliceInPlace()
|
| D | sort_util.h | 40 llvm::IRBuilder<>* b, const gpu::LaunchDimensions& launch_dimensions,
|
| D | dynamic_update_slice_util.h | 85 const gpu::LaunchDimensions& launch_dimensions, llvm::IRBuilder<>* b);
|
| D | sort_util.cc | 297 llvm::IRBuilder<>* b, const gpu::LaunchDimensions& launch_dimensions, in EmitSortInPlace()
|