Skip to content

GitLab

Commit e4ed3740 authored by ltqin's avatar ltqin
Browse files

modify some parameter

parent a76d2a3e
Hide whitespace changes
Inline Side-by-side
Showing with 29 additions and 112 deletions
host/driver_offline/include/device_convolution_backward_weight_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nhwk.hpp
View file @ e4ed3740
......@@ -76,7 +76,7 @@ void device_convolution_backward_weight_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nh
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 4;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#elif 1
#elif 0
// [M, N, K0, K1] = [128, 128, 4, 4] for fp32
constexpr index_t BlockSize = 256;
......@@ -91,105 +91,22 @@ void device_convolution_backward_weight_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nh
constexpr index_t MRepeat = 2;
constexpr index_t NRepeat = 2;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 2, 4>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmK1 = 1;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmK1 = 1;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 2, 4>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmK1 = 1;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 1;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#elif 0
// [M, N, K0, K1] = [256, 256, 4, 8] for fp16
constexpr index_t BlockSize = 256;
constexpr index_t GemmMPerBlock = 256;
constexpr index_t GemmNPerBlock = 256;
constexpr index_t GemmKPerBlock = 4;
constexpr index_t GemmMPerXDL = 32;
constexpr index_t GemmNPerXDL = 32;
constexpr index_t GemmK1 = 8;
constexpr index_t MRepeat = 4;
constexpr index_t NRepeat = 4;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 4, 8>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmK1 = 8;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 4, 8>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#elif 0
// [M, N, K0, K1] = [256, 128, 4, 8] for fp16
constexpr index_t BlockSize = 256;
constexpr index_t GemmMPerBlock = 256;
constexpr index_t GemmNPerBlock = 128;
constexpr index_t GemmKPerBlock = 4;
constexpr index_t GemmMPerXDL = 32;
constexpr index_t GemmNPerXDL = 32;
constexpr index_t GemmK1 = 8;
constexpr index_t MRepeat = 4;
constexpr index_t NRepeat = 2;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 4, 8>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 64, 1>;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 4, 2>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 32, 2>;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmM = 2;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmM = 2;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 2, 8>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 64, 1>;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 4, 2>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 32, 2>;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmN = 2;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmN = 2;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#elif 0
// [M, N, K0, K1] = [128, 256, 4, 8] for fp16
constexpr index_t BlockSize = 256;
constexpr index_t GemmMPerBlock = 128;
constexpr index_t GemmNPerBlock = 256;
constexpr index_t GemmKPerBlock = 4;
constexpr index_t GemmMPerXDL = 32;
constexpr index_t GemmNPerXDL = 32;
constexpr index_t GemmK1 = 8;
constexpr index_t MRepeat = 2;
constexpr index_t NRepeat = 4;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 2, 8>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmK1 = 8;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 4, 8>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 64, 1>;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#elif 0
// [M, N, K0, K1] = [128, 128, 4, 8] for fp16
#elif 1
// [M, N, K0, K1] = [128, 128, 4, 8]
constexpr index_t BlockSize = 256;
constexpr index_t GemmMPerBlock = 128;
......@@ -203,17 +120,17 @@ void device_convolution_backward_weight_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nh
constexpr index_t MRepeat = 2;
constexpr index_t NRepeat = 2;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 2, 8>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 64, 1>;
using GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1 = Sequence<1, 4, 4>;
using GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1 = Sequence<4, 32, 2>;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmABlockTransferSrcScalarPerVector_GemmM = 4;
constexpr index_t GemmABlockTransferDstScalarPerVector_GemmM = 4;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 2, 8>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 64, 1>;
using GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1 = Sequence<1, 4, 4>;
using GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1 = Sequence<4, 32, 2>;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmK1 = 8;
constexpr index_t GemmBBlockTransferSrcScalarPerVector_GemmN = 4;
constexpr index_t GemmBBlockTransferDstScalarPerVector_GemmN = 4;
constexpr index_t GemmCThreadTransferDstScalarPerVector = 1;
#endif
......@@ -294,19 +211,19 @@ void device_convolution_backward_weight_implicit_gemm_v4r4r4_xdlops_nhwc_kyxc_nh
NRepeat,
GemmABlockTransferThreadSliceLengths_GemmK0_GemmM_GemmK1,
GemmABlockTransferThreadClusterLengths_GemmK0_GemmM_GemmK1,
Sequence<1, 0, 2>,
Sequence<1, 0, 2>,
2,
GemmABlockTransferSrcScalarPerVector_GemmK1,
GemmABlockTransferDstScalarPerVector_GemmK1,
Sequence<0, 2, 1>,
Sequence<0, 2, 1>,
1,
GemmABlockTransferSrcScalarPerVector_GemmM,
GemmABlockTransferDstScalarPerVector_GemmM,
false, // don't move back src coordinate after threadwise copy
GemmBBlockTransferThreadSliceLengths_GemmK0_GemmN_GemmK1,
GemmBBlockTransferThreadClusterLengths_GemmK0_GemmN_GemmK1,
Sequence<1, 0, 2>,
Sequence<1, 0, 2>,
2,
GemmBBlockTransferSrcScalarPerVector_GemmK1,
GemmBBlockTransferDstScalarPerVector_GemmK1,
Sequence<0, 2, 1>,
Sequence<0, 2, 1>,
1,
GemmBBlockTransferSrcScalarPerVector_GemmN,
GemmBBlockTransferDstScalarPerVector_GemmN,
false, // don't move back src coordinate after threadwise copy
Sequence<2, 3, 0, 1, 7, 5, 4, 6>,
7,
......
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