| /aosp_15_r20/external/pytorch/aten/src/ATen/native/cuda/ |
| H A D | UniqueCub.cu | 83 int* inv_loc_ptr = static_cast<int*>(inv_loc.mutable_data_ptr()); in compute_unique()
98 inv_loc_out.mutable_data_ptr<int64_t>(), in compute_unique()
109 inverse_indices.mutable_data_ptr<int64_t>()); in compute_unique()
120 …cuda::cub::unique(data, data_out.mutable_data_ptr<scalar_t>(), length.mutable_data_ptr<int64_t>(),… in compute_unique()
126 data_out.mutable_data_ptr<scalar_t>(), in compute_unique()
127 counts.mutable_data_ptr<int64_t>(), in compute_unique()
128 length.mutable_data_ptr<int64_t>(), in compute_unique()
167 sorted.mutable_data_ptr<scalar_t>(), in operator ()()
177 sorted.mutable_data_ptr<scalar_t>(), in operator ()()
179 sorted_indices.mutable_data_ptr<int64_t>(), in operator ()()
[all …]
|
| H A D | EmbeddingBag.cu | 194 indices.const_data_ptr<index_t>(), sorted_indices.mutable_data_ptr<index_t>(), in embedding_bag_backward_cuda_sum_avg()
195 range.const_data_ptr<index_t>(), orig_indices.mutable_data_ptr<index_t>(), in embedding_bag_backward_cuda_sum_avg()
209 auto count_data = count.mutable_data_ptr<index_t>(); in embedding_bag_backward_cuda_sum_avg()
297 grad_weight.mutable_data_ptr<scalar_t>(), stride, numBags, in embedding_bag_backward_cuda_max()
404 weight.const_data_ptr<scalar_t>(), output.mutable_data_ptr<scalar_t>(), in _embedding_bag_cuda()
405 offset2bag.mutable_data_ptr<index_t>(), numIndices, numBags, featureSize, in _embedding_bag_cuda()
406 weight.stride(0), weight.stride(1), bag_size.mutable_data_ptr<index_t>(), in _embedding_bag_cuda()
407 max_indices.mutable_data_ptr<index_t>(), in _embedding_bag_cuda()
413 weight.const_data_ptr<scalar_t>(), output.mutable_data_ptr<scalar_t>(), in _embedding_bag_cuda()
414 offset2bag.mutable_data_ptr<index_t>(), numIndices, numBags, featureSize, in _embedding_bag_cuda()
[all …]
|
| H A D | EmbeddingBackwardKernel.cu | 236 int64_t *num_of_segments_ptr = num_of_segments_tensor.mutable_data_ptr<int64_t>(); in embedding_backward_cuda_kernel()
247 segment_offsets.mutable_data_ptr<index_t>(), in embedding_backward_cuda_kernel()
261 partials_per_segment.mutable_data_ptr<index_t>(), in embedding_backward_cuda_kernel()
275 partials_per_segment_offset.mutable_data_ptr<index_t>(), in embedding_backward_cuda_kernel()
280 … int64_t *num_of_partial_segments_ptr = num_of_partial_segments_tensor.mutable_data_ptr<int64_t>(); in embedding_backward_cuda_kernel()
294 partial_segment_offset.mutable_data_ptr<index_t>(), in embedding_backward_cuda_kernel()
330 … num_of_partial_segments_ptr, grad_weight_per_segment.mutable_data_ptr<partial_weight_t>(), in embedding_backward_cuda_kernel()
341 grad_weight_per_segment.mutable_data_ptr<partial_weight_t>(), in embedding_backward_cuda_kernel()
351 grad_weight.mutable_data_ptr<scalar_t>(), in embedding_backward_cuda_kernel()
|
| H A D | MultiMarginLoss.cu | 203 out.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
214 out.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
236 out.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
246 out.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
259 tmp_output.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
269 tmp_output.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_out()
337 grad_input.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_backward_out()
350 grad_input.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_backward_out()
372 grad_input.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_backward_out()
384 grad_input.mutable_data_ptr<scalar_t>(), in multi_margin_loss_cuda_backward_out()
|
| H A D | LegacyThrustHelpers.cu | 32 auto orig_data = device_ptr(orig_indices.mutable_data_ptr<int64_t>()); in index_put_with_sort_kernel_thrust_helper()
40 auto sorted_data = device_ptr(sorted_indices.mutable_data_ptr<int64_t>()); in index_put_with_sort_kernel_thrust_helper()
58 auto count_data = thrust::device_ptr<index_t>(count.mutable_data_ptr<index_t>()); in embedding_dense_backward_cuda_scan()
96 auto dummy_dev = thrust::device_ptr<index_t>(dummy.mutable_data_ptr<index_t>()); in embedding_backward_cuda_kernel_unique_by_key()
103 thrust::device_ptr<index_t>(segment_offsets.mutable_data_ptr<index_t>())); in embedding_backward_cuda_kernel_unique_by_key()
|
| H A D | MultiLabelMarginCriterion.cu | 234 output.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
237 is_target_.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
259 output_tmp.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
262 is_target_.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
285 output.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
288 is_target_.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_forward_out_cuda_template()
344 grad_input.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_backward_cuda_out_template()
373 grad_input.mutable_data_ptr<scalar_t>(), in multilabel_margin_loss_backward_cuda_out_template()
|
| H A D | WeightNorm.cu | 379 (w.mutable_data_ptr<scalar_t>(), in weight_norm_cuda()
380 norms.mutable_data_ptr<accscalar_t>(), in weight_norm_cuda()
410 (w.mutable_data_ptr<scalar_t>(), in weight_norm_cuda()
411 norms.mutable_data_ptr<accscalar_t>(), in weight_norm_cuda()
468 (grad_v.mutable_data_ptr<scalar_t>(), in weight_norm_backward_cuda()
469 grad_g.mutable_data_ptr<scalar_t>(), in weight_norm_backward_cuda()
501 (grad_v.mutable_data_ptr<scalar_t>(), in weight_norm_backward_cuda()
502 grad_g.mutable_data_ptr<scalar_t>(), in weight_norm_backward_cuda()
|
| H A D | Embedding.cu | 285 grad_weight.mutable_data_ptr<scalar_t>(), in embedding_dense_backward_cuda()
302 indices.const_data_ptr<index_t>(), sorted_indices.mutable_data_ptr<index_t>(), in embedding_dense_backward_cuda()
303 range.const_data_ptr<index_t>(), orig_indices.mutable_data_ptr<index_t>(), in embedding_dense_backward_cuda()
317 auto count_data = count.mutable_data_ptr<index_t>(); in embedding_dense_backward_cuda()
365 unique_indices.mutable_data_ptr<index_t>(), in embedding_renorm_cuda_()
366 num_unique_indices.mutable_data_ptr<int64_t>(), in embedding_renorm_cuda_()
382 self.mutable_data_ptr<scalar_t>(), in embedding_renorm_cuda_()
|
| H A D | Normalization.cuh | 1451 …accscalar_t* staging_data_ptr = grid.y > 1 ? staging_data.mutable_data_ptr<accscalar_t>() : nullpt… in batch_norm_stats_channels_last_cuda_template()
1452 int* semaphores_ptr = grid.y > 1 ? semaphores.mutable_data_ptr<int>() : nullptr; in batch_norm_stats_channels_last_cuda_template()
1456 out_mean.mutable_data_ptr<accscalar_t>(), in batch_norm_stats_channels_last_cuda_template()
1457 out_invstd.mutable_data_ptr<accscalar_t>(), in batch_norm_stats_channels_last_cuda_template()
1497 output.mutable_data_ptr<scalar_t>(), in batch_norm_elemt_channels_last_cuda_template()
1518 output.mutable_data_ptr<scalar_t>(), in batch_norm_elemt_channels_last_cuda_template()
1566 …accscalar_t* staging_data_ptr = grid.y > 1 ? staging_data.mutable_data_ptr<accscalar_t>() : nullpt… in batch_norm_backward_reduce_cuda_channels_last_template()
1567 int* semaphores_ptr = grid.y > 1 ? semaphores.mutable_data_ptr<int>() : nullptr; in batch_norm_backward_reduce_cuda_channels_last_template()
1574 sumn_dy.mutable_data_ptr<accscalar_t>(), in batch_norm_backward_reduce_cuda_channels_last_template()
1575 sum_dy_xmu.mutable_data_ptr<accscalar_t>(), in batch_norm_backward_reduce_cuda_channels_last_template()
[all …]
|
| H A D | NaiveDilatedConvolution.cu | 246 /* y=*/grad_bias.mutable_data_ptr<scalar_t>(), \ in slow_conv_dilated_all_cuda_template()
286 columns.mutable_data_ptr<scalar_t>()); in slow_conv_dilated_all_cuda_template()
302 /* C=*/output_n.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cuda_template()
327 /* C=*/columns.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cuda_template()
342 grad_input_n.mutable_data_ptr<scalar_t>()); in slow_conv_dilated_all_cuda_template()
358 columns.mutable_data_ptr<scalar_t>()); in slow_conv_dilated_all_cuda_template()
376 /* C=*/grad_weight.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cuda_template()
|
| /aosp_15_r20/external/executorch/backends/cadence/reference/operators/ |
| H A D | quantize_per_tensor.cpp | 35 uint8_t* out_data = out.mutable_data_ptr<uint8_t>(); in quantize_per_tensor_out()
39 int8_t* out_data = out.mutable_data_ptr<int8_t>(); in quantize_per_tensor_out()
45 uint16_t* out_data = out.mutable_data_ptr<uint16_t>(); in quantize_per_tensor_out()
49 int16_t* out_data = out.mutable_data_ptr<int16_t>(); in quantize_per_tensor_out()
53 int32_t* out_data = out.mutable_data_ptr<int32_t>(); in quantize_per_tensor_out()
|
| /aosp_15_r20/external/executorch/backends/cadence/hifi/operators/ |
| H A D | quantize_per_tensor.cpp | 36 uint8_t* out_data = out.mutable_data_ptr<uint8_t>(); in quantize_per_tensor_out()
40 int8_t* out_data = out.mutable_data_ptr<int8_t>(); in quantize_per_tensor_out()
44 int16_t* out_data = out.mutable_data_ptr<int16_t>(); in quantize_per_tensor_out()
50 uint16_t* out_data = out.mutable_data_ptr<uint16_t>(); in quantize_per_tensor_out()
54 int32_t* out_data = out.mutable_data_ptr<int32_t>(); in quantize_per_tensor_out()
|
| H A D | op_div.cpp | 94 float* a_data = a.mutable_data_ptr<float>(); in div_out()
95 float* b_data = b.mutable_data_ptr<float>(); in div_out()
96 float* out_data = out.mutable_data_ptr<float>(); in div_out()
215 float* a_data = a.mutable_data_ptr<float>(); in div_out_mode()
216 float* b_data = b.mutable_data_ptr<float>(); in div_out_mode()
217 float* out_data = out.mutable_data_ptr<float>(); in div_out_mode()
|
| /aosp_15_r20/external/executorch/extension/aten_util/test/ |
| H A D | make_aten_functor_from_et_functor_test.cpp | 30 out.mutable_data_ptr<int32_t>()[0] += 1; in add_1_out()
39 out.mutable_data_ptr<int64_t>()[0] += s1.value(); in add_optional_scalar_out()
42 out.mutable_data_ptr<int64_t>()[0] += s2.value(); in add_optional_scalar_out()
52 out.mutable_data_ptr<int64_t>()[0] += in add_optional_tensor_out()
53 s1.value().mutable_data_ptr<int64_t>()[0]; in add_optional_tensor_out()
56 out.mutable_data_ptr<int64_t>()[0] += in add_optional_tensor_out()
57 s2.value().mutable_data_ptr<int64_t>()[0]; in add_optional_tensor_out()
66 out.mutable_data_ptr<int64_t>()[0] += a[i]; in sum_arrayref_scalar_out()
75 out.mutable_data_ptr<int32_t>()[0] += a[i].const_data_ptr<int32_t>()[0]; in sum_arrayref_tensor_out()
86 out.mutable_data_ptr<int32_t>()[0] += in sum_arrayref_optional_tensor_out()
[all …]
|
| /aosp_15_r20/external/executorch/kernels/optimized/cpu/ |
| H A D | op_div.cpp | 83 out.mutable_data_ptr<CTYPE>(), in opt_div_out()
92 out.mutable_data_ptr<CTYPE>(), in opt_div_out()
117 out.mutable_data_ptr<CTYPE>(), in opt_div_out()
150 out.mutable_data_ptr<CTYPE>(), in opt_div_out()
158 out.mutable_data_ptr<CTYPE>(), in opt_div_out()
229 out.mutable_data_ptr<CTYPE>(), in opt_div_scalar_out()
251 out.mutable_data_ptr<CTYPE_OUT>(); in opt_div_scalar_out()
|
| H A D | op_sub.cpp | 123 out.mutable_data_ptr<CTYPE>(), in opt_sub_out()
131 out.mutable_data_ptr<CTYPE>(), in opt_sub_out()
160 out.mutable_data_ptr<CTYPE>(), in opt_sub_out()
197 out.mutable_data_ptr<CTYPE>(), in opt_sub_out()
205 out.mutable_data_ptr<CTYPE>(), in opt_sub_out()
287 out.mutable_data_ptr<CTYPE>(), in opt_sub_scalar_out()
308 CTYPE_OUT* out_data = out.mutable_data_ptr<CTYPE_OUT>(); in opt_sub_scalar_out()
|
| H A D | op_mul.cpp | 101 out.mutable_data_ptr<CTYPE>(), in handle_last_dim_broadcast()
170 out.mutable_data_ptr<CTYPE>(), in handle_broadcast_mul()
209 out.mutable_data_ptr<CTYPE>(), in opt_mul_out()
235 out.mutable_data_ptr<CTYPE>(), in opt_mul_out()
312 out.mutable_data_ptr<CTYPE>(), in opt_mul_scalar_out()
330 CTYPE_OUT* out_data = out.mutable_data_ptr<CTYPE_OUT>(); in opt_mul_scalar_out()
|
| /aosp_15_r20/external/executorch/runtime/executor/test/ |
| H A D | tensor_parser_test.cpp | 155 ASSERT_NE(t.mutable_data_ptr<float>()[0], 0.5); in TEST_F()
156 t.mutable_data_ptr<float>()[0] = 0.5; in TEST_F()
158 t.mutable_data_ptr<float>()[0], in TEST_F()
168 ASSERT_NE(t2.mutable_data_ptr<float>()[0], 0.5); in TEST_F()
174 t.mutable_data_ptr<float>()[0] = 0.5; in TEST_F()
183 ASSERT_NE(t3.mutable_data_ptr<float>()[0], 0.5); in TEST_F()
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/ao_sparse/quantized/cpu/ |
| H A D | qlinear_serialize.cpp | 26 vec.data(), vec.data() + vec.size(), t.mutable_data_ptr<UNDERLYING_DTYPE>()); in wrap_vector()
132 weight_values.mutable_data_ptr<uint8_t>(), in serialize()
173 w_scales_compact.mutable_data_ptr<float>()[0] = w_scales_data_ptr[0]; in serialize()
174 w_zero_points_compact.mutable_data_ptr<int8_t>()[0] = in serialize()
186 w_scales_compact.mutable_data_ptr<float>()); in serialize()
194 w_zero_points_compact.mutable_data_ptr<int8_t>(), in serialize()
|
| /aosp_15_r20/external/executorch/backends/qualcomm/runtime/ |
| H A D | QnnExecuTorchBackend.cpp | 119 args[i]->toTensor().mutable_data_ptr(), input_tensors[i]) != in execute()
132 void* mutable_data_ptr = in execute() local
133 args[output_index]->toTensor().mutable_data_ptr(); in execute()
134 if (qnn_manager->RegisterMem(mutable_data_ptr, output_tensor) != in execute()
136 output_tensor->FillDataBuffer(mutable_data_ptr, false /* copy_data */); in execute()
|
| /aosp_15_r20/external/executorch/runtime/core/exec_aten/testing_util/test/ |
| H A D | tensor_factory_test.cpp | 127 memcpy(expected.mutable_data_ptr<int32_t>(), data, sizeof(data)); in TEST_F()
156 memcpy(expected.mutable_data_ptr<float>(), data, sizeof(data)); in TEST_F()
186 memcpy(expected.mutable_data_ptr<bool>(), data, sizeof(data)); in TEST_F()
213 t1.mutable_data_ptr<int32_t>()[0] = 99; in TEST_F()
319 memcpy(expected.mutable_data_ptr<int>(), data, sizeof(data)); in TEST_F()
356 memcpy(expected.mutable_data_ptr<float>(), data, sizeof(data)); in TEST_F()
393 memcpy(expected.mutable_data_ptr<bool>(), data, sizeof(data)); in TEST_F()
460 t1.mutable_data_ptr<int32_t>()[0] = 99; in TEST_F()
513 memcpy(expected.mutable_data_ptr<int>(), data, sizeof(data)); in TEST_F()
627 memcpy(expected.mutable_data_ptr<int32_t>(), data, sizeof(data)); in TEST_F()
[all …]
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/nested/ |
| H A D | NestedTensorMatmul.cpp | 100 int64_t* self_new_sizes_ptr = self_new_sizes.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
103 int64_t* self_new_strides_ptr = self_new_strides.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
105 int64_t* self_new_offsets_ptr = self_new_offsets.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
109 int64_t* mat2_new_sizes_ptr = mat2_new_sizes.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
112 int64_t* mat2_new_strides_ptr = mat2_new_strides.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
114 int64_t* mat2_new_offsets_ptr = mat2_new_offsets.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
159 int64_t* out_new_offsets_ptr = out_new_offsets.mutable_data_ptr<int64_t>(); in matmul_with_bmm_nested()
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/ |
| H A D | NaiveDilatedConvolution.cpp | 269 columns.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
312 /* C=*/output_n.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
327 /* C=*/output_n.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
378 /* C=*/columns.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
393 /* C=*/columns.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
424 columns.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
468 /* C=*/grad_weight.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
483 /* C=*/grad_weight.mutable_data_ptr<scalar_t>(), in slow_conv_dilated_all_cpu_template()
|
| /aosp_15_r20/external/executorch/extension/kernel_util/test/ |
| H A D | make_boxed_from_unboxed_functor_test.cpp | 38 out.mutable_data_ptr<int32_t>()[0] = 1; in set_1_out()
50 out.mutable_data_ptr<int32_t>()[i] = sum; in add_tensor_out()
62 out.mutable_data_ptr<int32_t>()[0] += s1.value(); in add_optional_scalar_out()
65 out.mutable_data_ptr<int32_t>()[0] += s2.value(); in add_optional_scalar_out()
78 out.mutable_data_ptr<int32_t>()[j] += in add_optional_tensor_out()
|
| /aosp_15_r20/external/executorch/kernels/portable/cpu/ |
| H A D | op_linear_scratch_example.cpp | 89 scratch.mutable_data_ptr<scalar_t>() + (i * K + j); in linear_scratch_example()
109 scratch.mutable_data_ptr<scalar_t>() + (i * K + j); in linear_scratch_example()
110 scalar_t* out_ptr = out.mutable_data_ptr<scalar_t>() + (i * K + j); in linear_scratch_example()
111 scalar_t* bias_ptr = bias.value().mutable_data_ptr<scalar_t>() + j; in linear_scratch_example()
|