Home
last modified time | relevance | path

Searched refs:__restrict__ (Results 1 – 25 of 139) sorted by relevance

123456

/external/tensorflow/tensorflow/lite/kernels/internal/optimized/
Dsse_tensor_utils_impl.h23 #define __restrict__ __restrict macro
33 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
34 const int8_t* __restrict__ vectors,
35 const float* __restrict__ scaling_factors, int n_batch,
36 float* __restrict__ result);
41 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
42 const int8_t* __restrict__ vectors,
43 const float* __restrict__ scaling_factors, int n_batch, int32_t* scratch,
44 float* __restrict__ result, CpuBackendContext* context);
48 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
[all …]
Dneon_tensor_utils_impl.h22 #define __restrict__ __restrict macro
37 void NeonMatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix,
39 const int8_t* __restrict__ vectors,
42 float* __restrict__ result);
46 void NeonMatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix,
48 const int8_t* __restrict__ vectors,
51 float* __restrict__ result,
56 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
57 const int8_t* __restrict__ vectors, const float* scaling_factors,
58 int n_batch, float* __restrict__ result, const float* per_channel_scale,
[all …]
Dsse_tensor_utils.cc97 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in SseMatrixBatchVectorMultiplyAccumulateImpl()
98 const int8_t* __restrict__ vectors, in SseMatrixBatchVectorMultiplyAccumulateImpl()
99 const float* __restrict__ scaling_factors, int n_batch, in SseMatrixBatchVectorMultiplyAccumulateImpl()
100 float* __restrict__ result, const float* per_channel_scale, in SseMatrixBatchVectorMultiplyAccumulateImpl()
108 const int8_t* __restrict__ row_ptr = matrix + row * m_cols; in SseMatrixBatchVectorMultiplyAccumulateImpl()
209 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in SseMatrixBatchVectorMultiplyAccumulate()
210 const int8_t* __restrict__ vectors, in SseMatrixBatchVectorMultiplyAccumulate()
211 const float* __restrict__ scaling_factors, int n_batch, in SseMatrixBatchVectorMultiplyAccumulate()
212 float* __restrict__ result) { in SseMatrixBatchVectorMultiplyAccumulate()
220 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in SseMatrixBatchVectorMultiplyAccumulate()
[all …]
Dsse_tensor_utils.h45 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
46 const int8_t* __restrict__ vectors, in MatrixBatchVectorMultiplyAccumulate()
47 const float* __restrict__ scaling_factors, int n_batch, in MatrixBatchVectorMultiplyAccumulate()
48 float* __restrict__ result) { in MatrixBatchVectorMultiplyAccumulate()
54 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
55 const int8_t* __restrict__ vectors, const float* scaling_factors, in MatrixBatchVectorMultiplyAccumulate()
56 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate()
65 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
66 const int8_t* __restrict__ vectors, in MatrixBatchVectorMultiplyAccumulate()
67 const float* __restrict__ scaling_factors, int n_batch, in MatrixBatchVectorMultiplyAccumulate()
[all …]
Dneon_tensor_utils.h34 void MatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix, in MatrixBatchVectorMultiplyAccumulate()
36 const int8_t* __restrict__ vectors, in MatrixBatchVectorMultiplyAccumulate()
39 float* __restrict__ result) { in MatrixBatchVectorMultiplyAccumulate()
44 void MatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix, in MatrixBatchVectorMultiplyAccumulate()
46 const int8_t* __restrict__ vectors, in MatrixBatchVectorMultiplyAccumulate()
49 float* __restrict__ result, in MatrixBatchVectorMultiplyAccumulate()
56 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
57 const int8_t* __restrict__ vectors, const float* scaling_factors, in MatrixBatchVectorMultiplyAccumulate()
58 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate()
67 const float* __restrict__ matrix, const int32_t* __restrict__ segments, in SparseMatrixBatchVectorMultiplyAccumulate1x4()
[all …]
/external/tensorflow/tensorflow/lite/kernels/internal/
Dtensor_utils.h27 #define __restrict__ __restrict macro
43 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
44 const int8_t* __restrict__ vectors,
45 const float* __restrict__ scaling_factors, int n_batch,
46 int32_t* __restrict__ scratch, float* __restrict__ result,
47 CpuBackendContext* __restrict__ context);
51 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
52 const int8_t* __restrict__ vectors, const float* scaling_factors,
53 int n_batch, float* __restrict__ result, const float* per_channel_scale,
61 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
[all …]
Dtensor_utils_common.h22 #define __restrict__ __restrict macro
97 const float* __restrict__ matrix, const int32_t* __restrict__ segments,
98 const int32_t* __restrict__ indices, int m_rows, int m_cols,
99 const float* __restrict__ vector, int n_batch, float* __restrict__ result);
112 const float* __restrict__ matrix, const uint8_t* __restrict__ ledger,
113 int m_rows, int m_cols, const float* __restrict__ vector, int n_batch,
114 float* __restrict__ result);
124 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
125 const int8_t* __restrict__ vectors,
126 const float* __restrict__ scaling_factors, int n_batch,
[all …]
/external/tensorflow/tensorflow/lite/kernels/internal/reference/
Dportable_tensor_utils_impl.h22 #define __restrict__ __restrict macro
63 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
64 const int8_t* __restrict__ vectors, const float* scaling_factors,
65 int n_batch, float* __restrict__ result);
68 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
69 const int8_t* __restrict__ vectors, const float* scaling_factors,
70 int n_batch, float* __restrict__ result, const float* per_channel_scale,
75 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols,
76 const int8_t* __restrict__ vector, const float* scaling_factors,
77 int n_batch, int32_t* scratch, float* __restrict__ result,
[all …]
Dportable_tensor_utils.h21 #define __restrict__ __restrict macro
65 void MatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix, in MatrixBatchVectorMultiplyAccumulate()
67 const int8_t* __restrict__ vector, in MatrixBatchVectorMultiplyAccumulate()
70 float* __restrict__ result) { in MatrixBatchVectorMultiplyAccumulate()
76 const int8_t* __restrict__ matrix, const int m_rows, const int m_cols, in MatrixBatchVectorMultiplyAccumulate()
77 const int8_t* __restrict__ vectors, const float* scaling_factors, in MatrixBatchVectorMultiplyAccumulate()
78 int n_batch, float* __restrict__ result, const float* per_channel_scale, in MatrixBatchVectorMultiplyAccumulate()
87 void MatrixBatchVectorMultiplyAccumulate(const int8_t* __restrict__ matrix, in MatrixBatchVectorMultiplyAccumulate()
89 const int8_t* __restrict__ vector, in MatrixBatchVectorMultiplyAccumulate()
92 float* __restrict__ result, in MatrixBatchVectorMultiplyAccumulate()
[all …]
/external/tensorflow/tensorflow/core/kernels/
Dcompare_and_bitpack_op_gpu.cu.cc36 const T* __restrict__ threshold, in CompareAndBitpackKernel()
37 const T* __restrict__ input, in CompareAndBitpackKernel()
38 uint8* __restrict__ output) { in CompareAndBitpackKernel()
60 const int size, const bool* __restrict__ threshold, in CompareAndBitpackKernel()
61 const bool* __restrict__ input, uint8* __restrict__ output) { in CompareAndBitpackKernel()
81 const int size, const float* __restrict__ threshold, in CompareAndBitpackKernel()
82 const float* __restrict__ input, uint8* __restrict__ output) { in CompareAndBitpackKernel()
97 const int size, const double* __restrict__ threshold, in CompareAndBitpackKernel()
98 const double* __restrict__ input, uint8* __restrict__ output) { in CompareAndBitpackKernel()
Dcwise_op_clip_gpu.cu.cc28 const T *__restrict__ in0, in UnaryClipCustomKernel()
29 const T *__restrict__ in1, in UnaryClipCustomKernel()
30 const T *__restrict__ in2, in UnaryClipCustomKernel()
31 T *__restrict__ out) { in UnaryClipCustomKernel()
40 const T *__restrict__ in0, in BinaryRightClipCustomKernel()
41 const T *__restrict__ in1, in BinaryRightClipCustomKernel()
42 const T *__restrict__ in2, in BinaryRightClipCustomKernel()
43 T *__restrict__ out) { in BinaryRightClipCustomKernel()
52 const T *__restrict__ in0, in BinaryLeftClipCustomKernel()
53 const T *__restrict__ in1, in BinaryLeftClipCustomKernel()
[all …]
Dpopulation_count_op_gpu.cu.cc37 const T* __restrict__ input, in PopulationCountKernel()
38 uint8* __restrict__ output) { in PopulationCountKernel()
44 const int8* __restrict__ input, in PopulationCountKernel()
45 uint8* __restrict__ output) { in PopulationCountKernel()
54 const int16* __restrict__ input, in PopulationCountKernel()
55 uint8* __restrict__ output) { in PopulationCountKernel()
64 const int64* __restrict__ input, in PopulationCountKernel()
65 uint8* __restrict__ output) { in PopulationCountKernel()
Dscatter_functor_gpu.cu.h39 __device__ void operator()(T* __restrict__ dest, T src) const { *dest = src; }
44 __device__ void operator()(T* __restrict__ dest, T src) const {
51 __device__ void operator()(T* __restrict__ dest, T src) const {
58 __device__ void operator()(T* __restrict__ dest, T src) const {
65 __device__ void operator()(T* __restrict__ dest, T src) const {
72 __device__ void operator()(T* __restrict__ dest, T src) const {
79 __device__ void operator()(T* __restrict__ dest, T src) const {
85 __global__ void ScatterOpCustomKernel(T* __restrict__ params,
86 const T* __restrict__ updates,
87 const Index* __restrict__ indices,
[all …]
Dmaxpooling_op_gpu.cu.cc68 const int nthreads, const dtype* __restrict__ bottom_data, in MaxPoolForwardNCHW()
72 const int pad_l, dtype* __restrict__ top_data, int64* __restrict__ mask, in MaxPoolForwardNCHW()
112 const int nthreads, const int32* __restrict__ bottom_data, const int height, in MaxPoolForwardNoMaskKernel_NCHW_VECT_C()
116 int32* __restrict__ top_data) { in MaxPoolForwardNoMaskKernel_NCHW_VECT_C()
145 const int nthreads, const dtype* __restrict__ bottom_data, const int height, in MaxPoolForwardNHWC()
149 dtype* __restrict__ top_data, int64* __restrict__ mask, in MaxPoolForwardNHWC()
204 const dtype* __restrict__ top_diff, in MaxPoolBackward()
205 const int64* __restrict__ mask, in MaxPoolBackward()
207 dtype* __restrict__ bottom_diff, in MaxPoolBackward()
231 const int nthreads, const dtype* __restrict__ bottom_data, in MaxPoolGradBackwardNoMaskNCHW()
[all …]
Dbias_op_gpu.cu.cc54 __global__ void BiasNHWCKernel(int32 nthreads, const T* __restrict__ input, in BiasNHWCKernel()
55 const T* __restrict__ bias, in BiasNHWCKernel()
56 T* __restrict__ output, int32 bias_size) { in BiasNHWCKernel()
64 __global__ void BiasNCHWKernel(int32 nthreads, const T* __restrict__ input, in BiasNCHWKernel()
65 const T* __restrict__ bias, in BiasNCHWKernel()
66 T* __restrict__ output, int32 bias_size, in BiasNCHWKernel()
104 const T* __restrict__ output_backprop, in BiasGradNHWC_Naive()
105 T* __restrict__ bias_backprop, in BiasGradNHWC_Naive()
116 const T* __restrict__ output_backprop, in BiasGradNCHW_Naive()
117 T* __restrict__ bias_backprop, in BiasGradNCHW_Naive()
[all …]
Dsearchsorted_op_gpu.cu.cc34 __global__ void UpperBoundKernel(const T* __restrict__ sorted_inputs, in UpperBoundKernel()
36 int values_size, const T* __restrict__ values, in UpperBoundKernel()
37 OutType* __restrict__ outputs) { in UpperBoundKernel()
47 __global__ void LowerBoundKernel(const T* __restrict__ sorted_inputs, in LowerBoundKernel()
49 int values_size, const T* __restrict__ values, in LowerBoundKernel()
50 OutType* __restrict__ outputs) { in LowerBoundKernel()
Dpooling_ops_3d_gpu.cu.cc31 const int nthreads, const dtype* __restrict__ bottom_data, in MaxPoolGradBackwardNoMaskNCDHW()
32 const dtype* __restrict__ output_data, const int pooled_plane, in MaxPoolGradBackwardNoMaskNCDHW()
37 const int pad_l, const dtype* __restrict__ top_diff, in MaxPoolGradBackwardNoMaskNCDHW()
38 dtype* __restrict__ bottom_diff) { in MaxPoolGradBackwardNoMaskNCDHW()
82 const int nthreads, const dtype* __restrict__ bottom_data, in MaxPoolGradBackwardNoMaskNDHWC()
83 const dtype* __restrict__ output_data, const int pooled_plane, in MaxPoolGradBackwardNoMaskNDHWC()
88 const int pad_l, const dtype* __restrict__ top_diff, in MaxPoolGradBackwardNoMaskNDHWC()
89 dtype* __restrict__ bottom_diff) { in MaxPoolGradBackwardNoMaskNDHWC()
Ddilation_ops_gpu.cu.cc39 const int32 nthreads, const T* __restrict__ input_ptr, in DilationKernel()
40 const T* __restrict__ filter_ptr, int batch, int input_rows, int input_cols, in DilationKernel()
43 int rate_cols, int pad_top, int pad_left, T* __restrict__ output_ptr) { in DilationKernel()
78 const int32 nthreads, const T* __restrict__ input_ptr, in DilationBackpropInputKernel()
79 const T* __restrict__ filter_ptr, const T* __restrict__ out_backprop_ptr, in DilationBackpropInputKernel()
83 T* __restrict__ in_backprop_ptr) { in DilationBackpropInputKernel()
128 const int32 nthreads, const T* __restrict__ input_ptr, in DilationBackpropFilterKernel()
129 const T* __restrict__ filter_ptr, const T* __restrict__ out_backprop_ptr, in DilationBackpropFilterKernel()
133 T* __restrict__ filter_backprop_ptr) { in DilationBackpropFilterKernel()
Dfused_batch_norm_op.cu.cc135 int32 inner_dim_size, const T* __restrict__ in, in run()
136 const U* __restrict__ scale, in run()
137 const U* __restrict__ offset, in run()
138 const U* __restrict__ mean, in run()
139 const U* __restrict__ var, in run()
140 const T* __restrict__ side_input, float epsilon, in run()
141 T* __restrict__ out) { in run()
197 int32 inner_dim_size, const T* __restrict__ _in, in run()
198 const U* __restrict__ scale, in run()
199 const U* __restrict__ offset, in run()
[all …]
Droll_op_gpu.cu.cc34 const T* __restrict__ input, T* __restrict__ output, in RollKernel()
35 const int32* __restrict__ dim_size, in RollKernel()
36 const int32* __restrict__ threshold, in RollKernel()
37 const int64* __restrict__ dim_range) { in RollKernel()
/external/tensorflow/tensorflow/core/kernels/image/
Dresize_nearest_neighbor_op_gpu.cu.cc36 const int nthreads, const T* __restrict__ bottom_data, const int in_height, in ResizeNearestNeighborNHWC()
67 const int nthreads, const T* __restrict__ bottom_data, const int in_height, in LegacyResizeNearestNeighborNHWC()
70 T* __restrict__ top_data) { in LegacyResizeNearestNeighborNHWC()
96 const int nthreads, const T* __restrict__ top_diff, const int in_height, in ResizeNearestNeighborBackwardNHWC()
99 T* __restrict__ bottom_diff) { in ResizeNearestNeighborBackwardNHWC()
127 const int nthreads, const T* __restrict__ top_diff, const int in_height, in LegacyResizeNearestNeighborBackwardNHWC()
130 T* __restrict__ bottom_diff) { in LegacyResizeNearestNeighborBackwardNHWC()
176 void (*kernel)(const int nthreads, const T* __restrict__ bottom_data, in operator ()()
229 void (*kernel)(const int nthreads, const T* __restrict__ top_diff, in operator ()()
233 T* __restrict__ bottom_diff) = in operator ()()
Dcrop_and_resize_op_gpu.cu.cc41 const int32 nthreads, const T* __restrict__ image_ptr, in CropAndResizeKernel()
42 const float* __restrict__ boxes_ptr, const int32* __restrict__ box_ind_ptr, in CropAndResizeKernel()
45 float extrapolation_value, float* __restrict__ crops_ptr) { in CropAndResizeKernel()
134 const int32 nthreads, const float* __restrict__ grads_ptr, in CropAndResizeBackpropImageKernel()
135 const float* __restrict__ boxes_ptr, const int32* __restrict__ box_ind_ptr, in CropAndResizeBackpropImageKernel()
137 int crop_height, int crop_width, int depth, T* __restrict__ grads_image_ptr, in CropAndResizeBackpropImageKernel()
230 const int32 nthreads, const float* __restrict__ grads_ptr, in CropAndResizeBackpropBoxesKernel()
231 const T* __restrict__ image_ptr, const float* __restrict__ boxes_ptr, in CropAndResizeBackpropBoxesKernel()
232 const int32* __restrict__ box_ind_ptr, int num_boxes, int batch, in CropAndResizeBackpropBoxesKernel()
234 int depth, float* __restrict__ grads_boxes_ptr) { in CropAndResizeBackpropBoxesKernel()
Dresize_bilinear_op_gpu.cu.cc37 const int num_channel_threads, const T* __restrict__ images, in ResizeBilinearKernel_faster()
40 float* __restrict__ output) { in ResizeBilinearKernel_faster()
118 const int32 nthreads, const T* __restrict__ images, float height_scale, in ResizeBilinearKernel()
120 int out_height, int out_width, float* __restrict__ output) { in ResizeBilinearKernel()
169 const float* __restrict__ input_grad, in ResizeBilinearGradKernel()
174 T* __restrict__ output_grad) { in ResizeBilinearGradKernel()
234 const int32 nthreads, const float* __restrict__ input_grad, in ResizeBilinearDeterministicGradKernel()
238 float offset, T* __restrict__ output_grad) { in ResizeBilinearDeterministicGradKernel()
287 const int32 nthreads, const T* __restrict__ images, float height_scale, in LegacyResizeBilinearKernel()
289 int out_height, int out_width, float* __restrict__ output) { in LegacyResizeBilinearKernel()
[all …]
/external/tensorflow/tensorflow/core/kernels/linalg/
Dtridiagonal_matmul_op_gpu.cu.cc38 const Scalar* __restrict__ superdiag, in TridiagonalMatMulKernel()
39 const Scalar* __restrict__ maindiag, in TridiagonalMatMulKernel()
40 const Scalar* __restrict__ subdiag, in TridiagonalMatMulKernel()
41 const Scalar* __restrict__ rhs, in TridiagonalMatMulKernel()
42 Scalar* __restrict__ product) { in TridiagonalMatMulKernel()
/external/llvm-project/clang/test/CXX/over/over.match/over.match.funcs/
Dp4-0x.cpp24 void r() __restrict__; // expected-note {{'r' declared here}}
25 void cr() const __restrict__; // expected-note {{'cr' declared here}}
27 void vr() volatile __restrict__; // expected-note {{'vr' declared here}}
28 void cvr() const volatile __restrict__;
84 void test_diagnostics(const volatile X0 &__restrict__ cvr) { in test_diagnostics()

123456