Home
last modified time | relevance | path

Searched refs:num_cols (Results 1 – 25 of 96) sorted by relevance

1234

/external/libjpeg-turbo/simd/powerpc/
Djcgryext-altivec.c32 int pitch = img_width * RGB_PIXELSIZE, num_cols; in jsimd_rgb_gray_convert_altivec() local
68 for (num_cols = pitch; num_cols > 0; in jsimd_rgb_gray_convert_altivec()
69 num_cols -= RGB_PIXELSIZE * 16, inptr += RGB_PIXELSIZE * 16, in jsimd_rgb_gray_convert_altivec()
77 int bytes = num_cols + offset; in jsimd_rgb_gray_convert_altivec()
87 memcpy(tmpbuf, inptr, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_rgb_gray_convert_altivec()
116 if (num_cols < RGB_PIXELSIZE * 16 && (num_cols & 15)) { in jsimd_rgb_gray_convert_altivec()
118 memcpy(tmpbuf, inptr, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_rgb_gray_convert_altivec()
128 if (num_cols > 16) in jsimd_rgb_gray_convert_altivec()
130 if (num_cols > 32) in jsimd_rgb_gray_convert_altivec()
133 if (num_cols > 48) in jsimd_rgb_gray_convert_altivec()
[all …]
Djdcolext-altivec.c31 int pitch = out_width * RGB_PIXELSIZE, num_cols; in jsimd_ycc_rgb_convert_altivec() local
78 for (num_cols = pitch; num_cols > 0; in jsimd_ycc_rgb_convert_altivec()
79 num_cols -= RGB_PIXELSIZE * 16, outptr += RGB_PIXELSIZE * 16, in jsimd_ycc_rgb_convert_altivec()
203 int bytes = num_cols + offset; in jsimd_ycc_rgb_convert_altivec()
219 memcpy(outptr, tmpbuf, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_ycc_rgb_convert_altivec()
224 edgeh = vec_ld(min(num_cols - 1, RGB_PIXELSIZE * 16), outptr); in jsimd_ycc_rgb_convert_altivec()
250 if (num_cols < RGB_PIXELSIZE * 16 && (num_cols & 15)) { in jsimd_ycc_rgb_convert_altivec()
258 memcpy(outptr, tmpbuf, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_ycc_rgb_convert_altivec()
262 if (num_cols > 16) in jsimd_ycc_rgb_convert_altivec()
264 if (num_cols > 32) in jsimd_ycc_rgb_convert_altivec()
[all …]
Djccolext-altivec.c32 int pitch = img_width * RGB_PIXELSIZE, num_cols; in jsimd_rgb_ycc_convert_altivec() local
74 for (num_cols = pitch; num_cols > 0; in jsimd_rgb_ycc_convert_altivec()
75 num_cols -= RGB_PIXELSIZE * 16, inptr += RGB_PIXELSIZE * 16, in jsimd_rgb_ycc_convert_altivec()
83 int bytes = num_cols + offset; in jsimd_rgb_ycc_convert_altivec()
93 memcpy(tmpbuf, inptr, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_rgb_ycc_convert_altivec()
123 if (num_cols < RGB_PIXELSIZE * 16 && (num_cols & 15)) { in jsimd_rgb_ycc_convert_altivec()
125 memcpy(tmpbuf, inptr, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_rgb_ycc_convert_altivec()
135 if (num_cols > 16) in jsimd_rgb_ycc_convert_altivec()
137 if (num_cols > 32) in jsimd_rgb_ycc_convert_altivec()
140 if (num_cols > 48) in jsimd_rgb_ycc_convert_altivec()
Djdmrgext-altivec.c32 int pitch = output_width * RGB_PIXELSIZE, num_cols, yloop; in jsimd_h2v1_merged_upsample_altivec() local
86 for (num_cols = pitch; num_cols > 0; inptr1 += 16, inptr2 += 16) { in jsimd_h2v1_merged_upsample_altivec()
153 for (yloop = 0; yloop < 2 && num_cols > 0; yloop++, in jsimd_h2v1_merged_upsample_altivec()
154 num_cols -= RGB_PIXELSIZE * 16, in jsimd_h2v1_merged_upsample_altivec()
232 int bytes = num_cols + offset; in jsimd_h2v1_merged_upsample_altivec()
248 memcpy(outptr, tmpbuf, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_h2v1_merged_upsample_altivec()
253 edgeh = vec_ld(min(num_cols - 1, RGB_PIXELSIZE * 16), outptr); in jsimd_h2v1_merged_upsample_altivec()
279 if (num_cols < RGB_PIXELSIZE * 16 && (num_cols & 15)) { in jsimd_h2v1_merged_upsample_altivec()
287 memcpy(outptr, tmpbuf, min(num_cols, RGB_PIXELSIZE * 16)); in jsimd_h2v1_merged_upsample_altivec()
291 if (num_cols > 16) in jsimd_h2v1_merged_upsample_altivec()
[all …]
/external/tensorflow/tensorflow/python/kernel_tests/
Dcheckpoint_ops_test.py144 num_cols=self.old_num_cols)
159 num_cols=len(col_remapping))
174 num_cols=len(col_remapping))
189 num_cols=2)
209 num_cols=self.old_num_cols)
218 num_cols = 4
219 initializing_values = [42] * num_rows * num_cols
224 col_remapping=[-1] * num_cols,
227 num_cols=num_cols)
230 np.reshape(initializing_values, (num_rows, num_cols)),
[all …]
Dfractional_avg_pool_op_test.py152 num_cols = 6
153 tensor_shape = (1, num_rows, num_cols, 1)
180 for j in range(num_cols):
201 num_cols = 30
204 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
241 for num_cols in [10, 20, 50]:
242 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
257 num_cols = 50
258 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
278 num_cols = 50
[all …]
Dfractional_max_pool_op_test.py152 num_cols = 6
153 tensor_shape = (1, num_rows, num_cols, 1)
182 for j in range(num_cols):
203 num_cols = 30
206 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
219 num_cols = 30
222 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
238 for num_cols in [10, 20, 50]:
239 tensor_shape = (num_batches, num_rows, num_cols, num_channels)
254 num_cols = 50
[all …]
/external/tensorflow/tensorflow/python/data/experimental/benchmarks/
Dcsv_dataset_benchmark.py64 def _runBenchmark(self, dataset, num_cols, prefix): argument
88 name='%s_with_cols_%d' % (prefix, num_cols))
93 num_cols = self._num_cols[i]
94 kwargs = {'record_defaults': [[0.0]] * num_cols}
97 self._runBenchmark(dataset, num_cols, 'csv_float_map_decode_csv')
103 num_cols = self._num_cols[i]
104 kwargs = {'record_defaults': [['']] * num_cols}
107 self._runBenchmark(dataset, num_cols, 'csv_strings_map_decode_csv')
113 num_cols = self._num_cols[i]
114 kwargs = {'record_defaults': [[0.0]] * num_cols}
[all …]
/external/libjpeg-turbo/
Djdcol565.c29 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
58 num_cols--; in LOCAL()
60 for (col = 0; col < (num_cols >> 1); col++) { in LOCAL()
82 if (num_cols & 1) { in LOCAL()
108 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
139 num_cols--; in LOCAL()
141 for (col = 0; col < (num_cols >> 1); col++) { in LOCAL()
167 if (num_cols & 1) { in LOCAL()
192 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
211 num_cols--; in LOCAL()
[all …]
Djdcolext.c40 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
55 for (col = 0; col < num_cols; col++) { in LOCAL()
90 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
95 for (col = 0; col < num_cols; col++) { in LOCAL()
122 JDIMENSION num_cols = cinfo->output_width; in LOCAL() local
130 for (col = 0; col < num_cols; col++) { in LOCAL()
Djccolext.c42 JDIMENSION num_cols = cinfo->image_width; in LOCAL() local
50 for (col = 0; col < num_cols; col++) { in LOCAL()
96 JDIMENSION num_cols = cinfo->image_width; in LOCAL() local
102 for (col = 0; col < num_cols; col++) { in LOCAL()
129 JDIMENSION num_cols = cinfo->image_width; in LOCAL() local
137 for (col = 0; col < num_cols; col++) { in LOCAL()
/external/tensorflow/tensorflow/core/kernels/
Dtopk_op_gpu.h412 SegmentOffsetCreator(int num_cols) : num_cols_(num_cols) {}
422 ColumnIndexCreator(int num_cols) : num_cols_(num_cols) {}
434 int num_cols, int k,
445 DT_INT32, TensorShape({num_rows, num_cols}), &input_indices));
448 input_indices_t.generate(ColumnIndexCreator(num_cols));
453 segment_offsets_t(counting_iter, SegmentOffsetCreator(num_cols));
459 if (k == num_cols) {
466 DT_INT32, TensorShape({num_rows, num_cols}), &temp_indices));
468 TensorShape({num_rows, num_cols}),
481 /* num_items */ num_cols * num_rows,
[all …]
Dlu_op.cc85 const int64 num_cols = input.dim_size(input_rank - 1); in Compute() local
87 input_matrix_shape.AppendShape({num_rows, num_cols}); in Compute()
115 auto shard = [this, &input, &num_rows, &num_cols, &outputs, in Compute()
118 ComputeTensorSlice(context, i, input, num_rows, num_cols, outputs, in Compute()
129 const Tensor& input, int64 num_rows, int64 num_cols, in ComputeTensorSlice() argument
135 input.flat<Scalar>().data() + matrix_index * num_rows * num_cols, in ComputeTensorSlice()
136 num_rows, num_cols); in ComputeTensorSlice()
140 outputs[0]->flat<Scalar>().data() + matrix_index * num_rows * num_cols, in ComputeTensorSlice()
Dtopk_op.cc74 const int64 num_cols = input.dimension(1); in Compute() local
91 context, sorted_, k, input, num_rows, num_cols, values, indices); in Compute()
107 const int64 num_cols, typename TTypes<T, 2>::Tensor values, in Compute()
126 for (int c = 0; c < num_cols; ++c) { in Compute()
157 if (k == num_cols) { in Compute()
182 filter.reserve(num_cols); in Compute()
183 for (int32 c = 0; c < num_cols; ++c) { in Compute()
214 static_cast<double>(num_cols * in Compute()
216 const double sort_cost = (k == num_cols) ? base_cost : 4 * base_cost; in Compute()
253 const int64 num_cols, typename TTypes<T, 2>::Tensor values, \
Dreduction_gpu_kernels.cu.h232 T in, outT out, int num_rows, int num_cols, Op op,
242 if (num_cols == 1) {
251 if (row < num_rows && col < num_cols) {
252 sum = in[row * num_cols + col];
254 for (; col < num_cols; col += 32) {
255 sum = op(sum, in[row * num_cols + col]);
263 sum = WarpReduce(temp_storage).Reduce(sum, op, min(num_cols, 32));
299 T in, outT out, int num_rows, int num_cols, Op op,
302 int rows_per_warp = 32 / num_cols;
305 const int lane_row = lane / num_cols;
[all …]
Dsoftmax_op_gpu.cu.cc74 const int num_rows, const int num_cols, in GenerateNormalizedProb() argument
78 const int row = tid / num_cols; in GenerateNormalizedProb()
79 const int col = tid % num_cols; in GenerateNormalizedProb()
86 if (row < num_rows && col < num_cols) { in GenerateNormalizedProb()
100 const int num_cols) in SubtractAndExpFunctor()
101 : logits_(logits), max_logits_(max_logits), num_cols_(num_cols) {} in SubtractAndExpFunctor()
/external/libhevc/decoder/
Dihevcd_fmt_conv.c393 WORD32 num_rows, num_cols, src_strd, dst_strd; in ihevcd_fmt_conv_420sp_to_420sp() local
401 num_cols = wd; in ihevcd_fmt_conv_420sp_to_420sp()
408 memcpy(pu1_dst, pu1_src, num_cols); in ihevcd_fmt_conv_420sp_to_420sp()
418 num_cols = wd; in ihevcd_fmt_conv_420sp_to_420sp()
425 memcpy(pu1_dst, pu1_src, num_cols); in ihevcd_fmt_conv_420sp_to_420sp()
492 WORD32 num_rows, num_cols, src_strd, dst_strd; in ihevcd_fmt_conv_420sp_to_420sp_swap_uv() local
500 num_cols = wd; in ihevcd_fmt_conv_420sp_to_420sp_swap_uv()
507 memcpy(pu1_dst, pu1_src, num_cols); in ihevcd_fmt_conv_420sp_to_420sp_swap_uv()
517 num_cols = wd; in ihevcd_fmt_conv_420sp_to_420sp_swap_uv()
525 for(j = 0; j < num_cols; j += 2) in ihevcd_fmt_conv_420sp_to_420sp_swap_uv()
[all …]
/external/libavc/decoder/
Dih264d_format_conv.c385 WORD32 num_rows, num_cols, src_strd, dst_strd; in ih264d_fmt_conv_420sp_to_420sp() local
393 num_cols = wd; in ih264d_fmt_conv_420sp_to_420sp()
400 memcpy(pu1_dst, pu1_src, num_cols); in ih264d_fmt_conv_420sp_to_420sp()
410 num_cols = wd; in ih264d_fmt_conv_420sp_to_420sp()
417 memcpy(pu1_dst, pu1_src, num_cols); in ih264d_fmt_conv_420sp_to_420sp()
482 WORD32 num_rows, num_cols, src_strd, dst_strd; in ih264d_fmt_conv_420sp_to_420sp_swap_uv() local
490 num_cols = wd; in ih264d_fmt_conv_420sp_to_420sp_swap_uv()
497 memcpy(pu1_dst, pu1_src, num_cols); in ih264d_fmt_conv_420sp_to_420sp_swap_uv()
507 num_cols = wd; in ih264d_fmt_conv_420sp_to_420sp_swap_uv()
515 for(j = 0; j < num_cols; j += 2) in ih264d_fmt_conv_420sp_to_420sp_swap_uv()
[all …]
/external/libhevc/decoder/x86/
Dihevcd_fmt_conv_ssse3_intr.c63 WORD32 num_rows, num_cols, src_strd, dst_strd, cols, rows; in ihevcd_fmt_conv_420sp_to_420p_ssse3() local
76 num_cols = wd; in ihevcd_fmt_conv_420sp_to_420p_ssse3()
82 memcpy(pu1_dst, pu1_src, num_cols); in ihevcd_fmt_conv_420sp_to_420p_ssse3()
119 num_cols = wd >> 1; in ihevcd_fmt_conv_420sp_to_420p_ssse3()
126 if(num_cols > 15) in ihevcd_fmt_conv_420sp_to_420p_ssse3()
128 cols = num_cols >> 4; in ihevcd_fmt_conv_420sp_to_420p_ssse3()
246 num_cols &= 0x0F; in ihevcd_fmt_conv_420sp_to_420p_ssse3()
248 if(num_cols) in ihevcd_fmt_conv_420sp_to_420p_ssse3()
256 for(j = 0; j < num_cols; j++) in ihevcd_fmt_conv_420sp_to_420p_ssse3()
/external/libavc/encoder/
Dih264e_fmt_conv.c371 WORD32 num_rows, num_cols, src_strd, dst_strd; in ih264e_fmt_conv_420sp_to_420sp() local
379 num_cols = wd; in ih264e_fmt_conv_420sp_to_420sp()
386 memcpy(pu1_dst, pu1_src, num_cols); in ih264e_fmt_conv_420sp_to_420sp()
396 num_cols = wd; in ih264e_fmt_conv_420sp_to_420sp()
403 memcpy(pu1_dst, pu1_src, num_cols); in ih264e_fmt_conv_420sp_to_420sp()
423 WORD32 num_rows, num_cols, src_strd, dst_strd; in ih264e_fmt_conv_420sp_to_420sp_swap_uv() local
431 num_cols = wd; in ih264e_fmt_conv_420sp_to_420sp_swap_uv()
438 memcpy(pu1_dst, pu1_src, num_cols); in ih264e_fmt_conv_420sp_to_420sp_swap_uv()
448 num_cols = wd; in ih264e_fmt_conv_420sp_to_420sp_swap_uv()
456 for (j = 0; j < num_cols; j += 2) in ih264e_fmt_conv_420sp_to_420sp_swap_uv()
[all …]
/external/tensorflow/tensorflow/compiler/xla/service/gpu/
Dgemm_thunk.cc41 num_cols(matrix_num_cols), in MatrixDescriptor()
47 int64 num_cols; member
70 auto k = lhs_matrix.transpose ? lhs_matrix.num_rows : lhs_matrix.num_cols; in DoGemm()
76 output_matrix.num_cols, /*size of reduce dim=*/k, /*alpha=*/alpha, in DoGemm()
83 int64 lhs_stride = lhs_matrix.num_rows * lhs_matrix.num_cols; in DoGemm()
84 int64 rhs_stride = rhs_matrix.num_rows * rhs_matrix.num_cols; in DoGemm()
85 int64 output_stride = output_matrix.num_rows * output_matrix.num_cols; in DoGemm()
89 output_matrix.num_cols, /*size of reduce dim=*/k, in DoGemm()
135 auto k = lhs_matrix.transpose ? lhs_matrix.num_rows : lhs_matrix.num_cols; in DoGemmWithAlgorithm()
140 output_matrix.num_cols, /*size of reduce dim=*/k, in DoGemmWithAlgorithm()
[all …]
/external/eigen/unsupported/test/
Dcxx11_tensor_complex_cuda.cu80 const int num_cols = internal::random<int>(1024, 5*1024); in test_cuda_sum_reductions() local
82 Tensor<std::complex<float>, 2> in(num_rows, num_cols); in test_cuda_sum_reductions()
94 TensorMap<Tensor<std::complex<float>, 2> > in_gpu(gpu_in_ptr, num_rows, num_cols); in test_cuda_sum_reductions()
117 const int num_cols = internal::random<int>(1024, 5*1024); in test_cuda_product_reductions() local
119 Tensor<std::complex<float>, 2> in(num_rows, num_cols); in test_cuda_product_reductions()
131 TensorMap<Tensor<std::complex<float>, 2> > in_gpu(gpu_in_ptr, num_rows, num_cols); in test_cuda_product_reductions()
/external/libjpeg-turbo/simd/loongson/
Djdcolext-mmi.c97 int num_cols, col; in jsimd_ycc_rgb_convert_mmi() local
108 for (num_cols = out_width; num_cols > 0; num_cols -= 8, in jsimd_ycc_rgb_convert_mmi()
253 if (num_cols >= 8) { in jsimd_ycc_rgb_convert_mmi()
259 col = num_cols * 3; in jsimd_ycc_rgb_convert_mmi()
359 if (num_cols >= 8) { in jsimd_ycc_rgb_convert_mmi()
366 col = num_cols; in jsimd_ycc_rgb_convert_mmi()
/external/tensorflow/tensorflow/tools/graph_transforms/
Dfold_old_batch_norms.cc78 const int64 num_cols = mean.shape().dim_size(0); in GetScaleAndOffsetValues() local
79 TF_RETURN_IF_ERROR(ErrorIfNotVector(variance, "Variance", num_cols)); in GetScaleAndOffsetValues()
80 TF_RETURN_IF_ERROR(ErrorIfNotVector(beta, "Beta", num_cols)); in GetScaleAndOffsetValues()
81 TF_RETURN_IF_ERROR(ErrorIfNotVector(gamma, "gamma", num_cols)); in GetScaleAndOffsetValues()
83 scale_values->resize(num_cols); in GetScaleAndOffsetValues()
84 offset_values->resize(num_cols); in GetScaleAndOffsetValues()
88 for (int i = 0; i < num_cols; ++i) { in GetScaleAndOffsetValues()
94 for (int i = 0; i < num_cols; ++i) { in GetScaleAndOffsetValues()
99 for (int i = 0; i < num_cols; ++i) { in GetScaleAndOffsetValues()
/external/libmpeg2/common/
Dideint_utils.c260 WORD32 num_cols, num_rows; in ideint_pad_blk() local
266 num_cols = blk_wd + 4; in ideint_pad_blk()
274 num_cols -= 2; in ideint_pad_blk()
289 num_cols -= 2; in ideint_pad_blk()
296 memcpy(pu1_dst, pu1_src_top, num_cols); in ideint_pad_blk()
299 memcpy(pu1_dst, pu1_src_bot, num_cols); in ideint_pad_blk()

1234