Skip to content

GitLab

Commit d9eadda9 authored by Bartlomiej Kocot's avatar Bartlomiej Kocot
Browse files

Change gemm layout to 3d

parent 33b4b52c
Hide whitespace changes
Inline Side-by-side
Showing with 112 additions and 186 deletions
client_example/22_im2col_col2im/column_to_image.cpp
View file @ d9eadda9
......@@ -16,7 +16,7 @@
using InDataType = ck::half_t;
using OutDataType = ck::half_t;
using ImageLayout = ck::tensor_layout::convolution::GNHWC;
using ImageLayout = ck::tensor_layout::convolution::NHWGC;
static constexpr ck::index_t NumDimSpatial = 2;
static constexpr ck::index_t G = 2;
......@@ -56,7 +56,7 @@ int main()
// However, CK's API only accept length and stride with order of GNCHW
// Hence, we need to adjust the order of stride
std::array<ck::index_t, 5> image_strides{C, Hi * Wi * G * C, 1, Wi * G * C, G * C};
std::array<ck::index_t, 2> gemm_strides{Y * X * C, 1};
std::array<ck::index_t, 3> gemm_strides{Y * X * C, G * Y * X * C, 1};
std::array<ck::index_t, NumDimSpatial> filter_strides{1, 1};
std::array<ck::index_t, NumDimSpatial> filter_dilations{1, 1};
......
client_example/22_im2col_col2im/image_to_column.cpp
View file @ d9eadda9
......@@ -16,7 +16,7 @@
using InDataType = ck::half_t;
using OutDataType = ck::half_t;
using ImageLayout = ck::tensor_layout::convolution::GNHWC;
using ImageLayout = ck::tensor_layout::convolution::NHWGC;
static constexpr ck::index_t NumDimSpatial = 2;
static constexpr ck::index_t G = 2;
......@@ -52,11 +52,11 @@ int main()
std::array<ck::index_t, 2> wei_spatial_lengths{Y, X};
std::array<ck::index_t, 2> out_spatial_lengths{Ho, Wo};
// We have NHWGC in memory space (G is dummy)
// We have NHWGC in memory space
// However, CK's API only accept length and stride with order of GNCHW
// Hence, we need to adjust the order of stride
std::array<ck::index_t, 5> image_strides{C, Hi * Wi * G * C, 1, Wi * G * C, G * C};
std::array<ck::index_t, 2> gemm_strides{Y * X * C, 1};
std::array<ck::index_t, 3> gemm_strides{Y * X * C, G * Y * X * C, 1};
std::array<ck::index_t, NumDimSpatial> filter_strides{1, 1};
std::array<ck::index_t, NumDimSpatial> filter_dilations{1, 1};
......
example/52_im2col_col2im/column_to_image_f32.cpp
View file @ d9eadda9
......@@ -24,15 +24,14 @@ bool RunColumnToImage(const ExecutionConfig& config, const ck::utils::conv::Conv
const auto N = conv_params.N_;
const auto C = conv_params.C_;
const ck::index_t GNDoHoWo =
G * N *
ck::accumulate_n<ck::index_t>(
conv_params.output_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const ck::index_t NDoHoWo =
N * ck::accumulate_n<ck::index_t>(
conv_params.output_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const ck::index_t CZYX =
C * ck::accumulate_n<ck::index_t>(
conv_params.filter_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const auto in_desc = HostTensorDescriptor({GNDoHoWo, CZYX});
const auto in_desc = HostTensorDescriptor({G, NDoHoWo, CZYX});
const auto out_desc =
ck::utils::conv::make_input_host_tensor_descriptor_g_n_c_wis_packed<ImLayout>(conv_params);
......@@ -40,7 +39,7 @@ bool RunColumnToImage(const ExecutionConfig& config, const ck::utils::conv::Conv
std::array<ck::index_t, NDimSpatial> filter_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> output_spatial_lengths{};
std::array<ck::index_t, NDimSpatial + 3> image_g_n_c_wis_strides{};
std::array<ck::index_t, 2> gemm_m_k_strides{};
std::array<ck::index_t, 3> gemm_g_m_k_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_dilations{};
std::array<ck::index_t, NDimSpatial> input_left_pads{};
......@@ -51,7 +50,7 @@ bool RunColumnToImage(const ExecutionConfig& config, const ck::utils::conv::Conv
copy(conv_params.input_spatial_lengths_, input_spatial_lengths);
copy(conv_params.filter_spatial_lengths_, filter_spatial_lengths);
copy(conv_params.output_spatial_lengths_, output_spatial_lengths);
copy(in_desc.GetStrides(), gemm_m_k_strides);
copy(in_desc.GetStrides(), gemm_g_m_k_strides);
copy(out_desc.GetStrides(), image_g_n_c_wis_strides);
copy(conv_params.conv_filter_strides_, conv_filter_strides);
copy(conv_params.conv_filter_dilations_, conv_filter_dilations);
......@@ -94,7 +93,7 @@ bool RunColumnToImage(const ExecutionConfig& config, const ck::utils::conv::Conv
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......@@ -110,7 +109,7 @@ bool RunColumnToImage(const ExecutionConfig& config, const ck::utils::conv::Conv
}
float ave_time = invoker.Run(argument, StreamConfig{nullptr, config.time_kernel});
std::size_t num_btype = GNDoHoWo * CZYX * (sizeof(OutDataType) + sizeof(InDataType));
std::size_t num_btype = G * NDoHoWo * CZYX * (sizeof(OutDataType) + sizeof(InDataType));
float gb_per_sec = num_btype / 1.E6 / ave_time;
std::cout << "Perf: " << ave_time << " ms, " << gb_per_sec << " GB/s" << std::endl;
......
example/52_im2col_col2im/image_to_column_f32.cpp
View file @ d9eadda9
......@@ -24,23 +24,22 @@ bool RunImageToColumn(const ExecutionConfig& config, const ck::utils::conv::Conv
const auto N = conv_params.N_;
const auto C = conv_params.C_;
const ck::index_t GNDoHoWo =
G * N *
ck::accumulate_n<ck::index_t>(
conv_params.output_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const ck::index_t NDoHoWo =
N * ck::accumulate_n<ck::index_t>(
conv_params.output_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const ck::index_t CZYX =
C * ck::accumulate_n<ck::index_t>(
conv_params.filter_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const auto in_desc =
ck::utils::conv::make_input_host_tensor_descriptor_g_n_c_wis_packed<ImLayout>(conv_params);
const auto out_desc = HostTensorDescriptor({GNDoHoWo, CZYX});
const auto out_desc = HostTensorDescriptor({G, NDoHoWo, CZYX});
std::array<ck::index_t, NDimSpatial> input_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> filter_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> output_spatial_lengths{};
std::array<ck::index_t, NDimSpatial + 3> image_g_n_c_wis_strides{};
std::array<ck::index_t, 2> gemm_m_k_strides{};
std::array<ck::index_t, 3> gemm_m_k_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_dilations{};
std::array<ck::index_t, NDimSpatial> input_left_pads{};
......@@ -110,7 +109,7 @@ bool RunImageToColumn(const ExecutionConfig& config, const ck::utils::conv::Conv
}
float ave_time = invoker.Run(argument, StreamConfig{nullptr, config.time_kernel});
std::size_t num_btype = GNDoHoWo * CZYX * (sizeof(OutDataType) + sizeof(InDataType));
std::size_t num_btype = G * NDoHoWo * CZYX * (sizeof(OutDataType) + sizeof(InDataType));
float gb_per_sec = num_btype / 1.E6 / ave_time;
std::cout << "Perf: " << ave_time << " ms, " << gb_per_sec << " GB/s" << std::endl;
......
include/ck/tensor_operation/gpu/device/device_conv_tensor_rearrange.hpp
View file @ d9eadda9
......@@ -18,8 +18,8 @@ namespace device {
* the gemm problem (Image to Column) and
* conversion gemm form to the image (Column to Image).
* Supported layouts:
* [G, N, Di, Hi, Wi, C] <-> [G * N * Do * Ho * Wo, Z * Y * X * C]
* [N, Di, Hi, Wi, G, C] <-> [N * Do * Ho * Wo * G, Z * Y * X * C]
* [G, N, Di, Hi, Wi, C] <-> [G, N * Do * Ho * Wo, Z * Y * X * C]
* [N, Di, Hi, Wi, G, C] <-> [N * Do * Ho * Wo, G, Z * Y * X * C]
*
* \tparam NDimSpatial Number of spatial dimensions.
* \tparam ImageLayout Input Layout.
......@@ -47,7 +47,7 @@ struct DeviceConvTensorRearrange : public BaseOperator
* \param filter_spatial_lengths Filter spatial lengths.
* \param output_spatial_lengths Output spatial lengths.
* \param image_g_n_c_wis_strides Image strides in order [G, N, C, D, H, W].
* \param gemm_m_k_strides Gemm form strides.
* \param gemm_g_m_k_strides Gemm form strides.
* \param conv_filter_strides Convolution filter strides.
* \param conv_filter_dilations Convolution filter dilations.
* \param input_left_pads Convolution left pads.
......@@ -64,7 +64,7 @@ struct DeviceConvTensorRearrange : public BaseOperator
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......
include/ck/tensor_operation/gpu/device/impl/device_column_to_image_impl.hpp
View file @ d9eadda9
......@@ -25,9 +25,9 @@ namespace tensor_operation {
namespace device {
// Column to Image:
// input : gemm form [G * N * Do * Ho * Wo, Z * Y * X * C]
// input : gemm form [G, N * Do * Ho * Wo, Z * Y * X * C]
// output : input image [G, N, Di, Hi, Wi, C]
// input : gemm form [N * Do * Ho * Wo * G, Z * Y * X * C]
// input : gemm form [N * Do * Ho * Wo, G, Z * Y * X * C]
// output : input image [N, Di, Hi, Wi, G, C]
template <index_t NDimSpatial,
typename ImageLayout,
......@@ -96,13 +96,12 @@ struct DeviceColumnToImageImpl
// Make column form descriptor
static auto
MakeInputDescriptor_M_K(const ck::index_t G,
const ck::index_t N,
MakeInputDescriptor_M_K(const ck::index_t N,
const ck::index_t C,
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& independent_filters,
const std::array<index_t, NDimSpatial>& effs)
{
......@@ -112,26 +111,23 @@ struct DeviceColumnToImageImpl
C * ck::accumulate_n<index_t>(
filter_spatial_lengths.begin(), NDimSpatial, 1, std::multiplies<>());
const index_t AdditionalGroupStride = is_NSpatialGC ? G : 1;
const index_t NStride =
DoHoWo * gemm_m_k_strides[I0] * gemm_m_k_strides[I1] * AdditionalGroupStride;
const index_t NStride = DoHoWo * gemm_g_m_k_strides[I1] * gemm_g_m_k_strides[I2];
// Calculate the appropriate stride for each set of independent filters
// in each dimension
const index_t WStride = math::integer_divide_ceil(effs[XIdx], conv_filter_strides[XIdx]) *
gemm_m_k_strides[I0] * AdditionalGroupStride;
gemm_g_m_k_strides[I1];
const index_t HStride = math::integer_divide_ceil(effs[YIdx], conv_filter_strides[YIdx]) *
output_spatial_lengths[XIdx] * gemm_m_k_strides[I0] *
AdditionalGroupStride;
output_spatial_lengths[XIdx] * gemm_g_m_k_strides[I1];
const index_t DStride = math::integer_divide_ceil(effs[ZIdx], conv_filter_strides[ZIdx]) *
output_spatial_lengths[YIdx] * output_spatial_lengths[XIdx] *
gemm_m_k_strides[I0] * AdditionalGroupStride;
gemm_g_m_k_strides[I1];
// Create descriptor for independent filters in each dimension and
// then merge them into column form
if constexpr(NDimSpatial == 1)
{
const auto desc_gemm_form =
make_naive_tensor_descriptor(make_tuple(N, independent_filters[XIdx], CZYX),
make_tuple(NStride, WStride, gemm_m_k_strides[I1]));
make_tuple(NStride, WStride, gemm_g_m_k_strides[I2]));
const auto desc_gemm_form_merged_filters = transform_tensor_descriptor(
desc_gemm_form,
make_tuple(make_merge_transform(make_tuple(N, independent_filters[XIdx])),
......@@ -145,7 +141,7 @@ struct DeviceColumnToImageImpl
{
const auto desc_gemm_form = make_naive_tensor_descriptor(
make_tuple(N, independent_filters[YIdx], independent_filters[XIdx], CZYX),
make_tuple(NStride, HStride, WStride, gemm_m_k_strides[I1]));
make_tuple(NStride, HStride, WStride, gemm_g_m_k_strides[I2]));
const auto desc_gemm_form_merged_filters = transform_tensor_descriptor(
desc_gemm_form,
make_tuple(make_merge_transform(
......@@ -164,7 +160,7 @@ struct DeviceColumnToImageImpl
independent_filters[YIdx],
independent_filters[XIdx],
CZYX),
make_tuple(NStride, DStride, HStride, WStride, gemm_m_k_strides[I1]));
make_tuple(NStride, DStride, HStride, WStride, gemm_g_m_k_strides[I2]));
const auto desc_gemm_form_merged_filters = transform_tensor_descriptor(
desc_gemm_form,
make_tuple(make_merge_transform(make_tuple(N,
......@@ -259,7 +255,7 @@ struct DeviceColumnToImageImpl
}
using InputGridDesc =
remove_cvref_t<decltype(MakeInputDescriptor_M_K(1, 1, 1, {}, {}, {}, {}, {}, {}))>;
remove_cvref_t<decltype(MakeInputDescriptor_M_K(1, 1, {}, {}, {}, {}, {}, {}))>;
using OutputGridDesc = remove_cvref_t<decltype(MakeOutDescriptor_M_K(
1, 1, {}, {}, {}, {}, {}, {}, {}, {}, {}, {}))>;
......@@ -292,7 +288,7 @@ struct DeviceColumnToImageImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -308,21 +304,7 @@ struct DeviceColumnToImageImpl
input_left_pads_{input_left_pads},
input_right_pads_{input_right_pads}
{
using namespace tensor_layout::convolution;
if constexpr(is_NSpatialGC)
{
compute_ptr_offset_of_batch_.BatchStrideA_ =
gemm_m_k_strides[I0] * gemm_m_k_strides[I1];
}
else if constexpr(is_GNSpatialC)
{
const index_t NDoHoWo =
N * ck::accumulate_n<index_t>(
output_spatial_lengths.begin(), NDimSpatial, 1, std::multiplies<>());
compute_ptr_offset_of_batch_.BatchStrideA_ =
NDoHoWo * gemm_m_k_strides[I0] * gemm_m_k_strides[I1];
}
compute_ptr_offset_of_batch_.BatchStrideA_ = gemm_g_m_k_strides[I0];
compute_ptr_offset_of_batch_.BatchStrideC_ = image_g_n_c_wis_strides[I0];
const index_t x_eff =
......@@ -385,13 +367,12 @@ struct DeviceColumnToImageImpl
continue;
const auto in_grid_desc_m_k =
MakeInputDescriptor_M_K(G,
N,
MakeInputDescriptor_M_K(N,
C,
filter_spatial_lengths,
output_spatial_lengths,
conv_filter_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
independent_filters,
effs);
const auto out_grid_desc_m_k =
......@@ -421,13 +402,12 @@ struct DeviceColumnToImageImpl
const index_t z_offset_with_pad =
math::max(0, z_img_offset - input_left_pads[ZIdx]);
const index_t AdditionalGroupStride = is_NSpatialGC ? G : 1;
// Memory offsets to next set of independent filters,
// move to independent filters in each dimension
const index_t in_offset =
(x_idx + y_idx * output_spatial_lengths[XIdx] +
z_idx * output_spatial_lengths[YIdx] * output_spatial_lengths[XIdx]) *
gemm_m_k_strides[0] * AdditionalGroupStride;
gemm_g_m_k_strides[I1];
// Move to independent filters in appropriate dimensions
const index_t out_offset =
x_offset_with_pad * image_g_n_c_wis_strides[spatial_offset + XIdx] +
......@@ -583,7 +563,7 @@ struct DeviceColumnToImageImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -598,7 +578,7 @@ struct DeviceColumnToImageImpl
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......@@ -617,7 +597,7 @@ struct DeviceColumnToImageImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -632,7 +612,7 @@ struct DeviceColumnToImageImpl
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......
include/ck/tensor_operation/gpu/device/impl/device_image_to_column_impl.hpp
View file @ d9eadda9
......@@ -116,12 +116,11 @@ struct DeviceImageToColumnImpl
}
static auto
MakeOutDescriptor_M_K(const ck::index_t G,
const ck::index_t N,
MakeOutDescriptor_M_K(const ck::index_t N,
const ck::index_t C,
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, 2>& gemm_m_k_strides)
const std::array<index_t, 3>& gemm_g_m_k_strides)
{
const index_t NDoHoWo =
N * ck::accumulate_n<index_t>(
......@@ -130,24 +129,14 @@ struct DeviceImageToColumnImpl
C * ck::accumulate_n<index_t>(
filter_spatial_lengths.begin(), NDimSpatial, 1, std::multiplies<>());
if constexpr(is_NSpatialGC)
{
const auto desc_mraw_kraw = make_naive_tensor_descriptor(
make_tuple(NDoHoWo, CZYX),
make_tuple(gemm_m_k_strides[I0] * G, gemm_m_k_strides[I1]));
return matrix_padder.PadADescriptor_M_K(desc_mraw_kraw);
}
else if constexpr(is_GNSpatialC)
{
const auto desc_mraw_kraw = make_naive_tensor_descriptor(
make_tuple(NDoHoWo, CZYX), make_tuple(gemm_m_k_strides[I0], gemm_m_k_strides[I1]));
return matrix_padder.PadADescriptor_M_K(desc_mraw_kraw);
}
const auto desc_mraw_kraw = make_naive_tensor_descriptor(
make_tuple(NDoHoWo, CZYX), make_tuple(gemm_g_m_k_strides[I1], gemm_g_m_k_strides[I2]));
return matrix_padder.PadADescriptor_M_K(desc_mraw_kraw);
}
using InputGridDesc =
remove_cvref_t<decltype(MakeInputDescriptor_M_K(1, 1, {}, {}, {}, {}, {}, {}, {}, {}))>;
using OutputGridDesc = remove_cvref_t<decltype(MakeOutDescriptor_M_K(1, 1, 1, {}, {}, {}))>;
using OutputGridDesc = remove_cvref_t<decltype(MakeOutDescriptor_M_K(1, 1, {}, {}, {}))>;
using Block2ETileMap = remove_cvref_t<
decltype(BlockToCTileMap_M00_N0_M01Adapt<MPerBlock, KPerBlock, OutputGridDesc>(
......@@ -178,7 +167,7 @@ struct DeviceImageToColumnImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -207,22 +196,10 @@ struct DeviceImageToColumnImpl
input_right_pads);
out_grid_desc_m_k_ = MakeOutDescriptor_M_K(
G, N, C, filter_spatial_lengths, output_spatial_lengths, gemm_m_k_strides);
N, C, filter_spatial_lengths, output_spatial_lengths, gemm_g_m_k_strides);
compute_ptr_offset_of_batch_.BatchStrideA_ = image_g_n_c_wis_strides[I0];
if constexpr(is_NSpatialGC)
{
compute_ptr_offset_of_batch_.BatchStrideC_ =
gemm_m_k_strides[I0] * gemm_m_k_strides[I1];
}
else if constexpr(is_GNSpatialC)
{
const index_t NDoHoWo =
N * ck::accumulate_n<index_t>(
output_spatial_lengths.begin(), NDimSpatial, 1, std::multiplies<>());
compute_ptr_offset_of_batch_.BatchStrideC_ =
NDoHoWo * gemm_m_k_strides[I0] * gemm_m_k_strides[I1];
}
compute_ptr_offset_of_batch_.BatchStrideC_ = gemm_g_m_k_strides[I0];
}
void Print() const
......@@ -346,7 +323,7 @@ struct DeviceImageToColumnImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -361,7 +338,7 @@ struct DeviceImageToColumnImpl
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......@@ -380,7 +357,7 @@ struct DeviceImageToColumnImpl
const std::array<index_t, NDimSpatial>& filter_spatial_lengths,
const std::array<index_t, NDimSpatial>& output_spatial_lengths,
const std::array<index_t, NDimSpatial + 3>& image_g_n_c_wis_strides,
const std::array<index_t, 2>& gemm_m_k_strides,
const std::array<index_t, 3>& gemm_g_m_k_strides,
const std::array<index_t, NDimSpatial>& conv_filter_strides,
const std::array<index_t, NDimSpatial>& conv_filter_dilations,
const std::array<index_t, NDimSpatial>& input_left_pads,
......@@ -395,7 +372,7 @@ struct DeviceImageToColumnImpl
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......
library/include/ck/library/reference_tensor_operation/cpu/reference_column_to_image.hpp
View file @ d9eadda9
......@@ -92,9 +92,8 @@ struct ReferenceColumnToImage : public device::BaseOperator
float Run(const Argument& arg)
{
using namespace ck::tensor_layout::convolution;
if(!(arg.output_.GetNumOfDimension() == NDimSpatial + 3 &&
arg.input_.GetNumOfDimension() == 2))
arg.input_.GetNumOfDimension() == 3))
{
throw std::runtime_error("wrong! inconsistent dimension");
}
......@@ -111,14 +110,6 @@ struct ReferenceColumnToImage : public device::BaseOperator
{
index_t row = n * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNWC>)
{
row = g * N * Wo + n * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NWGC>)
{
row = n * Wo * G + wo * G + g;
}
for(index_t x = 0; x < arg.filter_spatial_lengths_[0]; ++x)
{
......@@ -131,7 +122,8 @@ struct ReferenceColumnToImage : public device::BaseOperator
if(wi >= 0 &&
ck::type_convert<std::size_t>(wi) < arg.output_.GetLengths()[3])
{
float v_in = ck::type_convert<float>(arg.input_(row, column));
float v_in =
ck::type_convert<float>(arg.input_(g, row, column));
float v_out = ck::type_convert<float>(arg.output_(g, n, c, wi));
arg.output_(g, n, c, wi) =
ck::type_convert<OutDataType>(v_in + v_out);
......@@ -156,16 +148,8 @@ struct ReferenceColumnToImage : public device::BaseOperator
{
for(index_t wo = 0; wo < Wo; ++wo)
{
index_t row = 0;
index_t row = n * Ho * Wo + ho * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNHWC>)
{
row = g * N * Ho * Wo + n * Ho * Wo + ho * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NHWGC>)
{
row = n * Ho * Wo * G + ho * Wo * G + wo * G + g;
}
for(index_t y = 0; y < arg.filter_spatial_lengths_[0]; ++y)
{
......@@ -192,7 +176,7 @@ struct ReferenceColumnToImage : public device::BaseOperator
arg.output_.GetLengths()[4])
{
float v_in =
ck::type_convert<float>(arg.input_(row, column));
ck::type_convert<float>(arg.input_(g, row, column));
float v_out = ck::type_convert<float>(
arg.output_(g, n, c, hi, wi));
arg.output_(g, n, c, hi, wi) =
......@@ -223,18 +207,8 @@ struct ReferenceColumnToImage : public device::BaseOperator
{
for(index_t wo = 0; wo < Wo; ++wo)
{
index_t row = 0;
index_t row = n * Do * Ho * Wo + d_o * Ho * Wo + ho * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNDHWC>)
{
row = g * N * Do * Ho * Wo + n * Do * Ho * Wo + d_o * Ho * Wo +
ho * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NDHWGC>)
{
row = n * Do * Ho * Wo * G + d_o * Ho * Wo * G + ho * Wo * G +
wo * G + g;
}
for(index_t z = 0; z < arg.filter_spatial_lengths_[0]; ++z)
{
......@@ -271,7 +245,7 @@ struct ReferenceColumnToImage : public device::BaseOperator
arg.output_.GetLengths()[5])
{
float v_in = ck::type_convert<float>(
arg.input_(row, column));
arg.input_(g, row, column));
float v_out = ck::type_convert<float>(
arg.output_(g, n, c, di, hi, wi));
arg.output_(g, n, c, di, hi, wi) =
......@@ -329,8 +303,9 @@ struct ReferenceColumnToImage : public device::BaseOperator
C * ck::accumulate_n<index_t>(
arg.filter_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
if(!(arg.input_.GetLengths()[0] == static_cast<std::size_t>(NDoHoWo) &&
arg.input_.GetLengths()[1] == static_cast<std::size_t>(CZYX)))
if(!(arg.input_.GetLengths()[0] == static_cast<std::size_t>(G) &&
arg.input_.GetLengths()[1] == static_cast<std::size_t>(NDoHoWo) &&
arg.input_.GetLengths()[2] == static_cast<std::size_t>(CZYX)))
{
return false;
}
......
library/include/ck/library/reference_tensor_operation/cpu/reference_image_to_column.hpp
View file @ d9eadda9
......@@ -92,9 +92,8 @@ struct ReferenceImageToColumn : public device::BaseOperator
float Run(const Argument& arg)
{
using namespace ck::tensor_layout::convolution;
if(!(arg.input_.GetNumOfDimension() == NDimSpatial + 3 &&
arg.output_.GetNumOfDimension() == 2))
arg.output_.GetNumOfDimension() == 3))
{
throw std::runtime_error("wrong! inconsistent dimension");
}
......@@ -107,16 +106,8 @@ struct ReferenceImageToColumn : public device::BaseOperator
{
const index_t Wo = arg.output_spatial_lengths_[0];
auto func = [&](auto g, auto n, auto wo) {
index_t row = 0;
index_t row = n * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNWC>)
{
row = g * N * Wo + n * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NWGC>)
{
row = n * Wo * G + wo * G + g;
}
for(index_t x = 0; x < arg.filter_spatial_lengths_[0]; ++x)
{
......@@ -129,8 +120,8 @@ struct ReferenceImageToColumn : public device::BaseOperator
if(wi >= 0 &&
ck::type_convert<std::size_t>(wi) < arg.input_.GetLengths()[3])
{
InDataType v_in = arg.input_(g, n, c, wi);
arg.output_(row, column) = ck::type_convert<OutDataType>(v_in);
InDataType v_in = arg.input_(g, n, c, wi);
arg.output_(g, row, column) = ck::type_convert<OutDataType>(v_in);
}
column++;
}
......@@ -147,16 +138,8 @@ struct ReferenceImageToColumn : public device::BaseOperator
const index_t Wo = arg.output_spatial_lengths_[1];
auto func = [&](auto g, auto n, auto ho, auto wo) {
index_t row = 0;
index_t row = n * Ho * Wo + ho * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNHWC>)
{
row = g * N * Ho * Wo + n * Ho * Wo + ho * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NHWGC>)
{
row = n * Ho * Wo * G + ho * Wo * G + wo * G + g;
}
for(index_t y = 0; y < arg.filter_spatial_lengths_[0]; ++y)
{
......@@ -178,8 +161,9 @@ struct ReferenceImageToColumn : public device::BaseOperator
wi >= 0 &&
ck::type_convert<std::size_t>(wi) < arg.input_.GetLengths()[4])
{
InDataType v_in = arg.input_(g, n, c, hi, wi);
arg.output_(row, column) = ck::type_convert<OutDataType>(v_in);
InDataType v_in = arg.input_(g, n, c, hi, wi);
arg.output_(g, row, column) =
ck::type_convert<OutDataType>(v_in);
}
column++;
}
......@@ -198,17 +182,8 @@ struct ReferenceImageToColumn : public device::BaseOperator
const index_t Wo = arg.output_spatial_lengths_[2];
auto func = [&](auto g, auto n, auto d_o, auto ho, auto wo) {
index_t row = 0;
index_t row = n * Do * Ho * Wo + d_o * Ho * Wo + ho * Wo + wo;
index_t column = 0;
if constexpr(std::is_same_v<ImageLayout, GNDHWC>)
{
row =
g * N * Do * Ho * Wo + n * Do * Ho * Wo + d_o * Ho * Wo + ho * Wo + wo;
}
else if constexpr(std::is_same_v<ImageLayout, NDHWGC>)
{
row = n * Do * Ho * Wo * G + d_o * Ho * Wo * G + ho * Wo * G + wo * G + g;
}
for(index_t z = 0; z < arg.filter_spatial_lengths_[0]; ++z)
{
......@@ -239,7 +214,7 @@ struct ReferenceImageToColumn : public device::BaseOperator
arg.input_.GetLengths()[5])
{
InDataType v_in = arg.input_(g, n, c, di, hi, wi);
arg.output_(row, column) =
arg.output_(g, row, column) =
ck::type_convert<OutDataType>(v_in);
}
column++;
......@@ -292,8 +267,9 @@ struct ReferenceImageToColumn : public device::BaseOperator
C * ck::accumulate_n<index_t>(
arg.filter_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
if(!(arg.output_.GetLengths()[0] == static_cast<std::size_t>(NDoHoWo) &&
arg.output_.GetLengths()[1] == static_cast<std::size_t>(CZYX)))
if(!(arg.output_.GetLengths()[0] == static_cast<std::size_t>(G) &&
arg.output_.GetLengths()[1] == static_cast<std::size_t>(NDoHoWo) &&
arg.output_.GetLengths()[2] == static_cast<std::size_t>(CZYX)))
{
return false;
}
......
profiler/include/profiler/profile_conv_tensor_rearrange_impl.hpp
View file @ d9eadda9
......@@ -93,6 +93,26 @@ static auto make_ref_op()
}
}
template <typename InputLayout>
static auto create_gemm_desc(const ck::index_t G, const ck::index_t NDoHoWo, const ck::index_t CZYX)
{
using namespace ck::tensor_layout::convolution;
if constexpr(std::is_same_v<InputLayout, GNWC> || std::is_same_v<InputLayout, GNHWC> ||
std::is_same_v<InputLayout, GNDHWC>)
{
return HostTensorDescriptor({G, NDoHoWo, CZYX});
}
else if constexpr(std::is_same_v<InputLayout, NWGC> || std::is_same_v<InputLayout, NHWGC> ||
std::is_same_v<InputLayout, NDHWGC>)
{
return HostTensorDescriptor({G, NDoHoWo, CZYX}, {CZYX, CZYX * G, 1});
}
else
{
throw std::runtime_error("Unsupported layout!");
}
}
template <index_t NDimSpatial,
typename InputLayout,
typename InputDataType,
......@@ -104,8 +124,8 @@ bool profile_conv_tensor_rearrange_impl(int do_verification,
bool time_kernel,
const ck::utils::conv::ConvParam& conv_param)
{
const ck::index_t GNDoHoWo =
conv_param.G_ * conv_param.N_ *
const ck::index_t NDoHoWo =
conv_param.N_ *
ck::accumulate_n<ck::index_t>(
conv_param.output_spatial_lengths_.begin(), NDimSpatial, 1, std::multiplies<>());
const ck::index_t CZYX =
......@@ -116,13 +136,13 @@ bool profile_conv_tensor_rearrange_impl(int do_verification,
const auto image_desc =
ck::utils::conv::make_input_host_tensor_descriptor_g_n_c_wis_packed<InputLayout>(
conv_param);
const auto gemm_desc = HostTensorDescriptor({GNDoHoWo, CZYX});
const auto gemm_desc = create_gemm_desc<InputLayout>(conv_param.G_, NDoHoWo, CZYX);
std::array<ck::index_t, NDimSpatial> input_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> filter_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> output_spatial_lengths{};
std::array<ck::index_t, NDimSpatial + 3> image_g_n_c_wis_strides{};
std::array<ck::index_t, 2> gemm_m_k_strides{};
std::array<ck::index_t, 3> gemm_g_m_k_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_dilations{};
std::array<ck::index_t, NDimSpatial> input_left_pads{};
......@@ -134,7 +154,7 @@ bool profile_conv_tensor_rearrange_impl(int do_verification,
copy(conv_param.filter_spatial_lengths_, filter_spatial_lengths);
copy(conv_param.output_spatial_lengths_, output_spatial_lengths);
copy(image_desc.GetStrides(), image_g_n_c_wis_strides);
copy(gemm_desc.GetStrides(), gemm_m_k_strides);
copy(gemm_desc.GetStrides(), gemm_g_m_k_strides);
copy(conv_param.conv_filter_strides_, conv_filter_strides);
copy(conv_param.conv_filter_dilations_, conv_filter_dilations);
copy(conv_param.input_left_pads_, input_left_pads);
......@@ -219,7 +239,7 @@ bool profile_conv_tensor_rearrange_impl(int do_verification,
filter_spatial_lengths,
output_spatial_lengths,
image_g_n_c_wis_strides,
gemm_m_k_strides,
gemm_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......@@ -235,7 +255,7 @@ bool profile_conv_tensor_rearrange_impl(int do_verification,
float avg_time =
invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, time_kernel});
std::size_t num_btype =
GNDoHoWo * CZYX * (sizeof(OutputDataType) + sizeof(InputDataType));
conv_param.G_ * NDoHoWo * CZYX * (sizeof(OutputDataType) + sizeof(InputDataType));
float gb_per_sec = num_btype / 1.E6 / avg_time;
std::cout << "Perf: " << std::setw(10) << avg_time << " ms, " << gb_per_sec << " GB/s, "
<< op_name << std::endl;
......
test/conv_tensor_rearrange/test_conv_tensor_rearrange_interface.cpp
View file @ d9eadda9
......@@ -71,13 +71,13 @@ class TestConvTensorRearrangeInterface : public ::testing::Test
const auto image_desc =
ck::utils::conv::make_input_host_tensor_descriptor_g_n_c_wis_packed<ImLayout>(
conv_param);
const auto gemm_desc = HostTensorDescriptor({NDoHoWo, CZYX});
const auto gemm_desc = HostTensorDescriptor({G, NDoHoWo, CZYX});
std::array<ck::index_t, NDimSpatial> input_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> filter_spatial_lengths{};
std::array<ck::index_t, NDimSpatial> output_spatial_lengths{};
std::array<ck::index_t, NDimSpatial + 3> input_g_n_c_wis_strides{};
std::array<ck::index_t, 2> output_m_k_strides{};
std::array<ck::index_t, 3> output_g_m_k_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_strides{};
std::array<ck::index_t, NDimSpatial> conv_filter_dilations{};
std::array<ck::index_t, NDimSpatial> input_left_pads{};
......@@ -89,7 +89,7 @@ class TestConvTensorRearrangeInterface : public ::testing::Test
copy(conv_param.filter_spatial_lengths_, filter_spatial_lengths);
copy(conv_param.output_spatial_lengths_, output_spatial_lengths);
copy(image_desc.GetStrides(), input_g_n_c_wis_strides);
copy(gemm_desc.GetStrides(), output_m_k_strides);
copy(gemm_desc.GetStrides(), output_g_m_k_strides);
copy(conv_param.conv_filter_strides_, conv_filter_strides);
copy(conv_param.conv_filter_dilations_, conv_filter_dilations);
copy(conv_param.input_left_pads_, input_left_pads);
......@@ -107,7 +107,7 @@ class TestConvTensorRearrangeInterface : public ::testing::Test
filter_spatial_lengths,
output_spatial_lengths,
input_g_n_c_wis_strides,
output_m_k_strides,
output_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......@@ -127,7 +127,7 @@ class TestConvTensorRearrangeInterface : public ::testing::Test
filter_spatial_lengths,
output_spatial_lengths,
input_g_n_c_wis_strides,
output_m_k_strides,
output_g_m_k_strides,
conv_filter_strides,
conv_filter_dilations,
input_left_pads,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment