/external/tensorflow/tensorflow/lite/delegates/gpu/common/ |
D | util_test.cc | 27 TEST(UtilTest, DivideRoundUp) { in TEST() argument 28 EXPECT_THAT(DivideRoundUp(0, 256), Eq(0)); in TEST() 29 EXPECT_THAT(DivideRoundUp(2u, 256), Eq(1)); in TEST() 30 EXPECT_THAT(DivideRoundUp(2, 256), Eq(1)); in TEST() 31 EXPECT_THAT(DivideRoundUp(255u, 256), Eq(1)); in TEST() 32 EXPECT_THAT(DivideRoundUp(255, 256), Eq(1)); in TEST() 33 EXPECT_THAT(DivideRoundUp(256u, 256), Eq(1)); in TEST() 34 EXPECT_THAT(DivideRoundUp(256, 256), Eq(1)); in TEST() 35 EXPECT_THAT(DivideRoundUp(257u, 256), Eq(2)); in TEST() 36 EXPECT_THAT(DivideRoundUp(257, 256), Eq(2)); in TEST()
|
D | util.h | 27 T DivideRoundUp(T n, N divisor) { in DivideRoundUp() function 34 inline uint3 DivideRoundUp(uint3 n, uint3 divisor) { in DivideRoundUp() function 35 return uint3(DivideRoundUp(n.x, divisor.x), DivideRoundUp(n.y, divisor.y), in DivideRoundUp() 36 DivideRoundUp(n.z, divisor.z)); in DivideRoundUp() 43 return DivideRoundUp(number, n) * n; in AlignByN()
|
D | convert.cc | 57 for (int p = 0; p < DivideRoundUp(shape.o, kPhwo4i4ChannelsInPlane); ++p) { in ConvertToPHWO4I4() 60 for (int c = 0; c < DivideRoundUp(shape.i, kPhwo4i4ChannelsInPlane); in ConvertToPHWO4I4() 118 DivideRoundUp(shape.o, 4)); in Get3DSizeForPHWO4I4() 135 const int dst_depth = DivideRoundUp(shape.o, 4); in ConvertToPHWO4I4() 136 const int src_depth = DivideRoundUp(shape.i, 4); in ConvertToPHWO4I4() 190 int32_t num_planes = DivideRoundUp(output_channels, kPiohw4ChannelsInPlane); in ConvertToPIOHW4() 243 int num_planes = DivideRoundUp(shape.c, kPhwc4ChannelsInPlane); in ConvertToPHWC4() 292 int num_planes = DivideRoundUp(shape.c, kPhwc4ChannelsInPlane); in ConvertToPHWC4Half() 418 int num_planes = DivideRoundUp(shape.c, kPhwc4ChannelsInPlane); in ConvertFromPHWC4() 460 int num_planes = DivideRoundUp(shape.c, kPhwc4ChannelsInPlane); in ConvertFromPHWC4Half()
|
/external/webrtc/rtc_base/numerics/ |
D | divide_round_unittest.cc | 21 static_assert(DivideRoundUp(5, 1) == 5, ""); in TEST() 22 static_assert(DivideRoundUp(5, 2) == 3, ""); in TEST() 26 EXPECT_EQ(DivideRoundUp(uint8_t{0}, 1), 0); in TEST() 27 EXPECT_EQ(DivideRoundUp(uint8_t{0}, 3), 0); in TEST() 28 EXPECT_EQ(DivideRoundUp(int{0}, 1), 0); in TEST() 29 EXPECT_EQ(DivideRoundUp(int{0}, 3), 0); in TEST() 33 EXPECT_EQ(DivideRoundUp(uint8_t{255}, 2), 128); in TEST() 34 EXPECT_EQ(DivideRoundUp(std::numeric_limits<int>::max(), 2), in TEST() 127 static_assert(std::is_same<decltype(DivideRoundUp(TypeParam{100}, int8_t{3})), in TYPED_TEST() 131 std::is_same<decltype(DivideRoundUp(TypeParam{100}, int16_t{3})), in TYPED_TEST() [all …]
|
/external/tensorflow/tensorflow/lite/delegates/gpu/common/task/ |
D | weights_conversion.h | 39 const int dst_slices = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsToOHWIOGroupI4O4() 40 const int src_slices = DivideRoundUp(weights.shape.i, 4); in RearrangeWeightsToOHWIOGroupI4O4() 41 const int dst_groups = DivideRoundUp(dst_slices, out_group_size); in RearrangeWeightsToOHWIOGroupI4O4() 75 const int dst_slices = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsToOHWIOGroupO4I4() 76 const int src_slices = DivideRoundUp(weights.shape.i, 4); in RearrangeWeightsToOHWIOGroupO4I4() 77 const int dst_groups = DivideRoundUp(dst_slices, out_group_size); in RearrangeWeightsToOHWIOGroupO4I4() 111 const int dst_slices = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsToODHWIOGroupI4O4() 112 const int src_slices = DivideRoundUp(weights.shape.i, 4); in RearrangeWeightsToODHWIOGroupI4O4() 113 const int dst_groups = DivideRoundUp(dst_slices, out_group_size); in RearrangeWeightsToODHWIOGroupI4O4() 149 const int dst_slices = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsToI4HWIOOGroupO4() [all …]
|
D | util.cc | 158 work_groups_count.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 159 work_groups_count.y = DivideRoundUp(grid_size.y, work_group_size.y); in GetWorkGroupsCount() 160 work_groups_count.z = DivideRoundUp(grid_size.z, work_group_size.z); in GetWorkGroupsCount()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/common/tasks/ |
D | conv_metal.cc | 44 const int dst_depth = DivideRoundUp(dst_channels, 4); in GetNumOutputSlices() 622 const int dst_slices = DivideRoundUp(dst_shape.c, 4); in GetGroupsCount() 624 int grid_x = DivideRoundUp(dst_shape.w, block_size.x); in GetGroupsCount() 625 int grid_y = DivideRoundUp(dst_shape.h, block_size.y); in GetGroupsCount() 626 int grid_z = DivideRoundUp(dst_slices, block_size.z); in GetGroupsCount() 628 return DivideRoundUp(grid_x, wg_size.x) * DivideRoundUp(grid_y, wg_size.y) * in GetGroupsCount() 629 DivideRoundUp(grid_z, wg_size.z); in GetGroupsCount() 634 const int dst_slices = DivideRoundUp(dst_shape.c, 4); in GetGroupsCountForLinearWH() 636 int grid_x = DivideRoundUp(dst_shape.w, block_size.x); in GetGroupsCountForLinearWH() 637 int grid_y = DivideRoundUp(dst_shape.h, block_size.y); in GetGroupsCountForLinearWH() [all …]
|
D | conv_constants.h | 36 const int dst_depth = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsForConvConstants() 37 const int src_depth = DivideRoundUp(weights.shape.i, 4); in RearrangeWeightsForConvConstants() 73 const int dst_depth = DivideRoundUp(weights.shape.o, 4); in RearrangeWeightsForConvConstantsDot() 74 const int src_depth = DivideRoundUp(weights.shape.i, 4); in RearrangeWeightsForConvConstantsDot() 111 const int src_depth = DivideRoundUp(weights.shape.i, 4); in UploadWeightsForConvConstants() 112 const int dst_depth = DivideRoundUp(weights.shape.o, 4); in UploadWeightsForConvConstants()
|
D | conv_buffer_1x1.cc | 303 const int dst_width_elements = DivideRoundUp( in GetGridSize() 306 DivideRoundUp(dst_width_elements, conv_params_.block_size.x); in GetGridSize() 308 DivideRoundUp(dst_[0]->Height(), conv_params_.block_size.y); in GetGridSize() 310 DivideRoundUp(dst_[0]->Slices(), conv_params_.block_size.z); in GetGridSize() 348 const int dst_depth = DivideRoundUp(attr.weights.shape.o, 4); in CreateConvBuffer1x1() 349 const int src_depth = DivideRoundUp(attr.weights.shape.i, 4); in CreateConvBuffer1x1() 366 const int dst_depth = DivideRoundUp(attr.weights.shape.o, 4); in CreateConvBuffer1x1() 367 const int src_depth = DivideRoundUp(attr.weights.shape.i, 4); in CreateConvBuffer1x1() 385 const int dst_depth = DivideRoundUp(attr.weights.shape.o, 4); in CreateConvBuffer1x1Wino4x4To6x6() 386 const int src_depth = DivideRoundUp(attr.weights.shape.i, 4); in CreateConvBuffer1x1Wino4x4To6x6() [all …]
|
D | conv_constants.cc | 115 const int out_z = DivideRoundUp(weights_shape.o, 4); in GenerateConvolutionConstantCode() 117 const int src_depth = DivideRoundUp(weights_shape.i, 4); in GenerateConvolutionConstantCode() 233 const int src_depth = DivideRoundUp(src_channels, 4); in IsDotConvBetter() 234 const int dst_depth = DivideRoundUp(dst_channels, 4); in IsDotConvBetter() 252 const int src_depth = DivideRoundUp(w_shape.i, 4); in IsConvConstantsSupported() 253 const int dst_depth = DivideRoundUp(w_shape.o, 4); in IsConvConstantsSupported() 262 const int flt4_registers = DivideRoundUp(w_shape.o, 4); in IsConvConstantsSupported()
|
D | fully_connected.h | 94 const int src_depth = DivideRoundUp(src_channels, 4); in RearrangeFCWeightsToOIO4I4() 96 const int dst_depth = DivideRoundUp(dst_channels, 4); in RearrangeFCWeightsToOIO4I4() 150 const int src_depth = DivideRoundUp(weights.shape.i, 4); in UploadWeights() 151 const int dst_depth = DivideRoundUp(weights.shape.o, 4); in UploadWeights()
|
D | winograd.cc | 194 int tiles_x = DivideRoundUp(new_width, 4); in GetGridSize() 195 int tiles_y = DivideRoundUp(new_height, 4); in GetGridSize() 204 int tiles_x = DivideRoundUp(new_width, 4); in BindArguments() 205 int tiles_y = DivideRoundUp(new_height, 4); in BindArguments() 448 const int tiles_x = DivideRoundUp( in BindArguments() 450 const int tiles_y = DivideRoundUp( in BindArguments() 667 const int tiles_x = DivideRoundUp(dst_[0]->Width(), 4); in BindArguments() 673 const int tiles_x = DivideRoundUp(dst_[0]->Width(), 4); in GetGridSize() 674 const int tiles_y = DivideRoundUp(dst_[0]->Height(), 4); in GetGridSize()
|
D | add.cc | 29 int dst_depth = DivideRoundUp(dst_channels, 4); in CreateAdd() 30 int src0_depth = DivideRoundUp(channels[0], 4); in CreateAdd()
|
D | depthwise_conv.h | 40 const int dst_depth = DivideRoundUp(dst_channels, 4); in RearrangeWeightsForDWConv2D() 71 const int dst_slices = DivideRoundUp(dst_channels, 4); in UploadWeightsForDWConv2D() 111 const int dst_slices = DivideRoundUp(dst_channels, 4); in RearrangeWeightsForDWConv3D() 145 const int dst_slices = DivideRoundUp(dst_channels, 4); in UploadWeightsForDWConv3D()
|
D | convolution_transposed.h | 120 AlignByN(DivideRoundUp(weights.shape.o, 4), block_size_.w); in UploadWeights() 121 const int src_depth = DivideRoundUp(weights.shape.i, 4); in UploadWeights() 144 AlignByN(DivideRoundUp(weights.shape.o, 4), block_size_.w); in UploadWeights() 145 const int src_depth = DivideRoundUp(weights.shape.i, 4); in UploadWeights()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/gl/kernels/ |
D | conv.cc | 62 {"src_depth", DivideRoundUp(weights.i, 4)}, in GenerateCode() 78 {"src_depth", DivideRoundUp(weights.i, 4)}, in GenerateCode() 191 DivideRoundUp(static_cast<int>(ctx.input_shapes[0][3]), 4)}, in GenerateCode() 196 MakeReadonlyObject(uint3(4, DivideRoundUp(attr.weights.shape.i, 4), in GenerateCode() 197 DivideRoundUp(attr.weights.shape.o, 4)), in GenerateCode() 234 auto dst_depth = DivideRoundUp(ctx.output_shapes[0][3], 4); in GenerateCode() 275 DivideRoundUp(ctx.output_shapes[0][3], 4)), in GenerateCode()
|
D | prelu.cc | 72 DivideRoundUp(static_cast<int>(ctx.output_shapes[0][3]), in GenerateCode() 104 DivideRoundUp(static_cast<int>(ctx.output_shapes[0][3]), 4)); in GenerateCode() 119 DivideRoundUp(static_cast<int>(ctx.output_shapes[0][3]), in GenerateCode() 139 DivideRoundUp(static_cast<int>(ctx.output_shapes[0][3]), in GenerateCode()
|
D | mul.cc | 112 DivideRoundUp(static_cast<int>(ctx.input_shapes[0][3]), 4)), in GenerateMultiplyScalarCode() 129 DivideRoundUp(static_cast<int>(ctx.input_shapes[0][3]), 4)), in GenerateMultiplyScalarCode() 137 DivideRoundUp(static_cast<int>(ctx.input_shapes[0][3]), 4)), in GenerateMultiplyScalarCode()
|
D | fully_connected.cc | 42 const int src_depth = DivideRoundUp(attr.weights.shape.i, 4); in GenerateCode() 43 const int dst_depth = DivideRoundUp(attr.weights.shape.o, 4); in GenerateCode()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/metal/ |
D | compute_task.cc | 42 work_groups_count.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 47 wgs.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 48 wgs.y = DivideRoundUp(grid_size.y, work_group_size.y); in GetWorkGroupsCount() 54 wgs.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 55 wgs.y = DivideRoundUp(grid_size.y, work_group_size.y); in GetWorkGroupsCount() 56 wgs.z = DivideRoundUp(grid_size.z, work_group_size.z); in GetWorkGroupsCount()
|
D | buffer_convert.mm | 25 using ::tflite::gpu::DivideRoundUp; 115 int slices = DivideRoundUp(shape.c, 4); 116 int groups_x = DivideRoundUp(shape.w * shape.b, group_size.width); 117 int groups_y = DivideRoundUp(shape.h, group_size.height); 118 int groups_z = DivideRoundUp(slices, group_size.depth);
|
/external/tensorflow/tensorflow/lite/delegates/gpu/cl/ |
D | cl_operation.cc | 27 work_groups_count.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 32 wgs.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 33 wgs.y = DivideRoundUp(grid_size.y, work_group_size.y); in GetWorkGroupsCount() 39 wgs.x = DivideRoundUp(grid_size.x, work_group_size.x); in GetWorkGroupsCount() 40 wgs.y = DivideRoundUp(grid_size.y, work_group_size.y); in GetWorkGroupsCount() 41 wgs.z = DivideRoundUp(grid_size.z, work_group_size.z); in GetWorkGroupsCount()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/gl/ |
D | object.h | 156 DivideRoundUp(static_cast<uint32_t>(data.size()), 4U), data); in MakeReadonlyObject() 161 DivideRoundUp(static_cast<uint32_t>(data.size()), 4U), data); in MakeReadonlyTexture() 166 DivideRoundUp(static_cast<uint32_t>(data.size()), 4U), data); in MakeReadonlyBuffer() 175 size.z = shape.b * DivideRoundUp(shape.c, 4); in GetPHWC4Size()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/common/tasks/special/ |
D | fc_fc_add.h | 70 const int src_depth = DivideRoundUp(src_channels, 4); in RearrangeFCWeightsToOIO4I4() 72 const int dst_depth = DivideRoundUp(dst_channels, 4); in RearrangeFCWeightsToOIO4I4() 127 const int src_depth = DivideRoundUp(weights.shape.i, 4); in UploadWeights() 128 const int dst_depth = DivideRoundUp(weights.shape.o, 4); in UploadWeights()
|
/external/tensorflow/tensorflow/lite/delegates/gpu/gl/converters/ |
D | bhwc_to_phwc4.cc | 91 uint3 workload = uint3(shape.w, shape.h, DivideRoundUp(shape.c, 4)); in Convert() 92 uint3 num_workgroups = DivideRoundUp(workload, workgroup_size_); in Convert()
|