Skip to content

GitLab

Commit 1fb77ae6 authored by Jing Zhang's avatar Jing Zhang
Browse files

fixed incorrect results due to typo

parent 64705e7d
Hide whitespace changes
Inline Side-by-side
Showing with 18 additions and 18 deletions
host/driver_offline/include/device_convolution_add_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -111,7 +111,7 @@ void device_convolution_add_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0
constexpr auto WoPerBlock = 32;
constexpr auto E1 = C0 * 9;
constexpr auto E2 = 1;
constexpr auto E2 = C1 / InWeiVectorSize;
constexpr auto K2 = 2;
constexpr auto E1PerBlock = C0;
......
host/driver_offline/include/device_convolution_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -104,7 +104,7 @@ void device_convolution_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1
constexpr index_t WoPerBlock = 32;
constexpr index_t E1 = C0 * Y * X;
constexpr index_t E2 = 1;
constexpr index_t E2 = C1 / InWeiVectorSize;
constexpr index_t K2 = 2;
constexpr index_t E1PerBlock = C0;
......@@ -126,9 +126,9 @@ void device_convolution_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1
#endif
const auto in_n_c0_hi_wi_c1_desc =
make_naive_tensor_descriptor_packed(make_tuple(N, C0, Hi, Wi, C1));
make_naive_tensor_descriptor_packed(make_tuple(N, C0, Hi, Wi, E2));
const auto wei_k_c0_y_x_c1_desc =
make_naive_tensor_descriptor_packed(make_tuple(K, C0, Y, X, C1));
make_naive_tensor_descriptor_packed(make_tuple(K, C0, Y, X, E2));
const auto out_n_k0_ho_wo_k1_desc =
make_naive_tensor_descriptor_packed(make_tuple(N, K0, Ho, Wo, K1));
......
host/driver_offline/include/device_convolution_maxpool_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -113,7 +113,7 @@ void device_convolution_maxpool_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1
constexpr auto WoPerBlock = 32;
constexpr auto E1 = C0 * 9;
constexpr auto E2 = 1;
constexpr auto E2 = C1 / InWeiVectorSize;
constexpr auto K2 = 2;
constexpr auto E1PerBlock = C0;
......
host/driver_offline/include/driver_convolution_add_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -128,7 +128,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
// input tensor
const auto in_n_c0_hip_wip_e2_global_desc = transform_tensor_descriptor(
make_naive_tensor_descriptor_packed(make_tuple(K, C0, Hi, Wi, E2)),
make_naive_tensor_descriptor_packed(make_tuple(N, C0, Hi, Wi, E2)),
make_tuple(make_pass_through_transform(N),
make_pass_through_transform(C0),
make_pad_transform(Hi, InLeftPadH, InRightPadH),
......@@ -149,7 +149,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(
Sequence<0>{}, Sequence<1>{}, Sequence<2, 3>{}, Sequence<4, 5>{}, Sequence<6>{}));
const auto b_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
const auto in_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
in_n_c0_y_ho_x_wo_e2_global_desc,
make_tuple(make_merge_transform(make_tuple(C0, Y, X)),
make_pass_through_transform(N),
......@@ -161,7 +161,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}, Sequence<4>{}));
const auto b_e0_e1_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
b_e_n_ho_wo_e2_grid_desc,
in_e_n_ho_wo_e2_grid_desc,
make_tuple(make_unmerge_transform(make_tuple(E0, E1)),
make_pass_through_transform(N),
make_pass_through_transform(Hop),
......
host/driver_offline/include/driver_convolution_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -123,7 +123,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
// input tensor
const auto in_n_c0_hip_wip_e2_global_desc = transform_tensor_descriptor(
make_naive_tensor_descriptor_packed(make_tuple(K, C0, Hi, Wi, E2)),
make_naive_tensor_descriptor_packed(make_tuple(N, C0, Hi, Wi, E2)),
make_tuple(make_pass_through_transform(N),
make_pass_through_transform(C0),
make_pad_transform(Hi, InLeftPadH, InRightPadH),
......@@ -144,7 +144,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(
Sequence<0>{}, Sequence<1>{}, Sequence<2, 3>{}, Sequence<4, 5>{}, Sequence<6>{}));
const auto b_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
const auto in_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
in_n_c0_y_ho_x_wo_e2_global_desc,
make_tuple(make_merge_transform(make_tuple(C0, Y, X)),
make_pass_through_transform(N),
......@@ -156,7 +156,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}, Sequence<4>{}));
const auto b_e0_e1_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
b_e_n_ho_wo_e2_grid_desc,
in_e_n_ho_wo_e2_grid_desc,
make_tuple(make_unmerge_transform(make_tuple(E0, E1)),
make_pass_through_transform(N),
make_pass_through_transform(Hop),
......
host/driver_offline/include/driver_convolution_maxpool_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 1fb77ae6
......@@ -129,7 +129,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
// input tensor
const auto in_n_c0_hip_wip_e2_global_desc = transform_tensor_descriptor(
make_naive_tensor_descriptor_packed(make_tuple(K, C0, Hi, Wi, E2)),
make_naive_tensor_descriptor_packed(make_tuple(N, C0, Hi, Wi, E2)),
make_tuple(make_pass_through_transform(N),
make_pass_through_transform(C0),
make_pad_transform(Hi, InLeftPadH, InRightPadH),
......@@ -150,7 +150,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(
Sequence<0>{}, Sequence<1>{}, Sequence<2, 3>{}, Sequence<4, 5>{}, Sequence<6>{}));
const auto b_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
const auto in_e_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
in_n_c0_y_ho_x_wo_e2_global_desc,
make_tuple(make_merge_transform(make_tuple(C0, Y, X)),
make_pass_through_transform(N),
......@@ -162,7 +162,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}, Sequence<4>{}));
const auto b_e0_e1_n_ho_wo_e2_grid_desc = transform_tensor_descriptor(
b_e_n_ho_wo_e2_grid_desc,
in_e_n_ho_wo_e2_grid_desc,
make_tuple(make_unmerge_transform(make_tuple(E0, E1)),
make_pass_through_transform(N),
make_pass_through_transform(Hop),
......
host/driver_offline/src/conv_fwd_driver_offline_nchwc.cpp
View file @ 1fb77ae6
......@@ -105,14 +105,14 @@ int main(int argc, char* argv[])
constexpr auto C1 = Number<8>{};
constexpr auto K0 = Number<1>{};
constexpr auto K1 = Number<4>{};
#elif 0
#elif 1
constexpr auto N = Number<1>{};
constexpr auto Hi = Number<1080>{};
constexpr auto Wi = Number<1920>{};
constexpr auto Y = Number<3>{};
constexpr auto X = Number<3>{};
constexpr auto C0 = Number<2>{};
constexpr auto C1 = Number<8>{};
constexpr auto C0 = Number<3>{};
constexpr auto C1 = Number<4>{};
constexpr auto K0 = Number<2>{};
constexpr auto K1 = Number<8>{};
#elif 0
......@@ -135,7 +135,7 @@ int main(int argc, char* argv[])
constexpr auto C1 = Number<8>{};
constexpr auto K0 = Number<2>{};
constexpr auto K1 = Number<8>{};
#elif 1
#elif 0
constexpr auto N = Number<128>{};
constexpr auto Hi = Number<270>{};
constexpr auto Wi = Number<480>{};
......
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