| /external/tensorflow/tensorflow/core/kernels/ |
| D | spacetobatch_op.cc | 55 const int block_dims = orig_block_shape.dim_size(0); in SpaceToBatchOpCompute() local
56 if (orig_input_tensor.dims() < 1 + block_dims) { in SpaceToBatchOpCompute()
57 return errors::InvalidArgument("input rank should be >= ", 1 + block_dims, in SpaceToBatchOpCompute()
62 block_dims == orig_paddings.dim_size(0) && in SpaceToBatchOpCompute()
64 return errors::InvalidArgument("paddings should have shape [", block_dims, in SpaceToBatchOpCompute()
79 for (; removed_prefix_block_dims < block_dims; ++removed_prefix_block_dims) { in SpaceToBatchOpCompute()
90 for (; removed_suffix_block_dims < block_dims - removed_prefix_block_dims; in SpaceToBatchOpCompute()
92 const int dim = block_dims - 1 - removed_suffix_block_dims; in SpaceToBatchOpCompute()
101 for (int block_dim = 0; block_dim < block_dims; ++block_dim) { in SpaceToBatchOpCompute()
110 block_dims - removed_prefix_block_dims - removed_suffix_block_dims; in SpaceToBatchOpCompute()
[all …]
|
| D | batchtospace_op.cc | 53 const int block_dims = orig_block_shape.dim_size(0); in BatchToSpaceOpCompute() local
55 context, orig_input_tensor.dims() >= 1 + block_dims, in BatchToSpaceOpCompute()
56 errors::InvalidArgument("input rank should be >= ", 1 + block_dims, in BatchToSpaceOpCompute()
61 block_dims == orig_crops.dim_size(0) && in BatchToSpaceOpCompute()
63 errors::InvalidArgument("crops should have shape [", block_dims, in BatchToSpaceOpCompute()
76 for (; removed_prefix_block_dims < block_dims; ++removed_prefix_block_dims) { in BatchToSpaceOpCompute()
87 for (; removed_suffix_block_dims < block_dims - removed_prefix_block_dims; in BatchToSpaceOpCompute()
89 const int dim = block_dims - 1 - removed_suffix_block_dims; in BatchToSpaceOpCompute()
98 for (int block_dim = 0; block_dim < block_dims; ++block_dim) { in BatchToSpaceOpCompute()
114 block_dims - removed_prefix_block_dims - removed_suffix_block_dims; in BatchToSpaceOpCompute()
[all …]
|
| /external/tensorflow/tensorflow/stream_executor/ |
| D | trace_listener.h | 50 const BlockDim& block_dims, in LaunchSubmit() argument
|
| D | stream_executor_pimpl.cc | 466 const BlockDim &block_dims, in Launch() argument
469 SubmitTrace(&TraceListener::LaunchSubmit, stream, thread_dims, block_dims, in Launch()
472 return implementation_->Launch(stream, thread_dims, block_dims, kernel, args); in Launch()
|
| D | stream_executor_internal.h | 191 const BlockDim &block_dims, const KernelBase &k, in Launch() argument
|
| D | stream_executor_pimpl.h | 475 const BlockDim &block_dims, const KernelBase &kernel,
|
| D | stream.h | 185 Stream &ThenLaunch(ThreadDim thread_dims, BlockDim block_dims,
2119 inline Stream &Stream::ThenLaunch(ThreadDim thread_dims, BlockDim block_dims, in ThenLaunch() argument
2134 parent_->Launch(this, thread_dims, block_dims, kernel, kernel_args) in ThenLaunch()
|
| /external/tensorflow/tensorflow/stream_executor/rocm/ |
| D | rocm_gpu_executor.cc | 297 const BlockDim& block_dims, in Launch() argument
311 VlogOccupancyInfo(kernel, thread_dims, block_dims); in Launch()
343 GetGpuContext(stream), hipfunc, block_dims.x, block_dims.y, block_dims.z, in Launch()
421 const BlockDim& block_dims) { in VlogOccupancyInfo() argument
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| D | cuda_gpu_executor.cc | 401 const BlockDim& block_dims, in Launch() argument
415 VlogOccupancyInfo(kernel, thread_dims, block_dims); in Launch()
431 context_, cufunc, block_dims.x, block_dims.y, block_dims.z, thread_dims.x, in Launch()
441 const BlockDim& block_dims) { in VlogOccupancyInfo() argument
|
| /external/tensorflow/tensorflow/stream_executor/gpu/ |
| D | gpu_executor.h | 73 const BlockDim& block_dims, const KernelBase& k,
268 const BlockDim& block_dims);
|
| /external/tensorflow/tensorflow/compiler/tests/ |
| D | randomized_tests.cc | 1232 std::vector<int64> block_dims = in TEST_F() local
1240 input_dims[1 + i] = block_dims[i]; in TEST_F()
1266 std::vector<int64> block_dims = RandomDims(1, 3, 0, 5); in TEST_F() local
1267 int num_block_dims = block_dims.size(); in TEST_F()
1270 RandomDims(block_dims.size(), block_dims.size(), 0, 4); in TEST_F()
1275 input_dims[0] *= block_dims[i]; in TEST_F()
1298 std::vector<int32>(block_dims.begin(), block_dims.end()))) in TEST_F()
2958 std::vector<int64> block_dims = RandomDims(4, 4, 0, 5); in TEST_F() local
2965 input_dims[1 + i] = block_dims[i] * block_size; in TEST_F()
2997 std::vector<int64> block_dims = RandomDims(1, 3, 0, 5); in TEST_F() local
[all …]
|
| /external/tensorflow/tensorflow/python/ops/linalg/ |
| D | linear_operator_util.py | 568 def split_arg_into_blocks(block_dims, block_dims_fn, arg, axis=-1): argument
584 block_sizes = [dim.value for dim in block_dims]
|
| /external/tensorflow/tensorflow/stream_executor/host/ |
| D | host_gpu_executor.h | 58 const BlockDim &block_dims, const KernelBase &kernel, in Launch() argument
|
| /external/tensorflow/tensorflow/compiler/xla/service/interpreter/ |
| D | executor.h | 65 const BlockDim &block_dims, const KernelBase &kernel, in Launch() argument
|
| /external/llvm-project/polly/lib/External/ppcg/ |
| D | cuda.c | 393 const char *block_dims[] = { "blockIdx.x", "blockIdx.y" }; in print_kernel_iterators() local
399 print_iterators(out, type, kernel->block_ids, block_dims); in print_kernel_iterators()
|