ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-17 11:30:02 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
5ba0f64087d00f21eb3b992cef623bdb11077cd0
composable_kernel/src/include
History
Chao Liu 5ba0f64087 adding fp16 direct that reads pre-vectorized data 
[ROCm/composable_kernel commit: 4f0fc72e91]
2019-03-18 15:03:17 -05:00
..
Array.hip.hpp
refactoring ConstantTensorDescriptor
2019-03-17 03:22:41 -05:00
blockwise_2d_tensor_op.hip.hpp
adding fp16 direct that reads pre-vectorized data
2019-03-18 15:03:17 -05:00
blockwise_4d_tensor_op.hip.hpp
adding fp16 direct that reads pre-vectorized data
2019-03-18 15:03:17 -05:00
blockwise_direct_convolution.hip.hpp
clean up
2019-03-02 17:27:37 -06:00
blockwise_gemm.hip.hpp
device_implicit_gemm_convolution_1_chwn_csrk_khwn: use tensor copy (instead of pointwise) for writing output, 3x3 increased from 78% to 84%, 5x5 from 80% to 84%
2019-02-19 11:47:46 -06:00
common.hip.hpp
adding fp16 direct that reads pre-vectorized data
2019-03-18 15:03:17 -05:00
config.h.in
refactor
2019-03-04 17:09:20 -06:00
constant_integral.hip.hpp
refactoring ConstantTensorDescriptor
2019-03-17 03:22:41 -05:00
ConstantMatrixDescriptor.hip.hpp
change file extension to hip.hpp and hip.cpp
2019-02-15 02:13:21 -06:00
ConstantTensorDescriptor.hip.hpp
refactoring ConstantTensorDescriptor
2019-03-17 03:22:41 -05:00
conv_common.hip.hpp
clean up
2019-03-02 17:27:37 -06:00
device.hpp
change file extension to hip.hpp and hip.cpp
2019-02-15 02:13:21 -06:00
functional.hip.hpp
refactoring ConstantTensorDescriptor
2019-03-17 03:22:41 -05:00
gridwise_direct_convolution_1.hip.hpp
clean up
2019-03-02 17:27:37 -06:00
gridwise_direct_convolution_2_nchw_kcyx_nkhw.hip.hpp
adding fp16 direct that reads pre-vectorized data
2019-03-18 15:03:17 -05:00
gridwise_direct_convolution_2_vectorized_nchw_kcyx_nkhw.hip.hpp
adding fp16 direct that reads pre-vectorized data
2019-03-18 15:03:17 -05:00
gridwise_implicit_gemm_convolution_1_chwn_cyxk_khwn_padded.hip.hpp
refactor
2019-03-09 13:39:24 -06:00
gridwise_implicit_gemm_convolution_1_chwn_cyxk_khwn.hip.hpp
refactor
2019-03-09 13:39:24 -06:00
gridwise_implicit_gemm_convolution_2_chwn_cyxk_khwn_lds_double_buffer.hip.hpp
refactor
2019-03-16 10:50:46 -05:00
gridwise_implicit_gemm_convolution_2_chwn_cyxk_khwn.hip.hpp
refactor
2019-03-16 10:50:46 -05:00
Sequence.hip.hpp
refactoring ConstantTensorDescriptor
2019-03-17 03:22:41 -05:00
tensor.hpp
refactor
2019-03-09 12:52:16 -06:00
threadwise_2d_tensor_op.hip.hpp
change file extension to hip.hpp and hip.cpp
2019-02-15 02:13:21 -06:00
threadwise_4d_tensor_op.hip.hpp
device_implicit_gemm_convolution_1_chwn_csrk_khwn: use tensor copy (instead of pointwise) for writing output, 3x3 increased from 78% to 84%, 5x5 from 80% to 84%
2019-02-19 11:47:46 -06:00
threadwise_direct_convolution.hip.hpp
refactor
2019-03-09 12:52:16 -06:00
threadwise_gemm.hip.hpp
change file extension to hip.hpp and hip.cpp
2019-02-15 02:13:21 -06:00
threadwise_nd_tensor_op.hip.hpp
refactor
2019-03-06 12:34:31 -06:00