Commits · 49998ba2b01d3bf5317205097fc199af9385e50a · gaoqiong / composable_kernel

03 Jan, 2023 2 commits
- Add test · 49998ba2
  rocking authored Jan 03, 2023
  
  49998ba2
- Add ckProfiler · 762a1ada
  rocking authored Jan 03, 2023
  
  762a1ada
30 Dec, 2022 1 commit
- Add gemm + layernorm instance · f3115bcd
  rocking authored Dec 30, 2022
  
  f3115bcd
28 Dec, 2022 4 commits
- 1. Declare e inside the host_gemm_layernorm() · bcdacc1f
  rocking authored Dec 28, 2022
```
2. Prevent implicit cast in reference code
```
  bcdacc1f
- Sync the variable with previous if · 29ad7a36
  rocking authored Dec 28, 2022
  
  29ad7a36
- Fix Typo · 29df52fb
  rocking authored Dec 28, 2022
  
  29df52fb
- Add missing index · b2c458ca
  rocking authored Dec 28, 2022
  
  b2c458ca
27 Dec, 2022 2 commits
- Merge the mean and var threadwise copy · 69b925b9
  rocking authored Dec 27, 2022
  
  69b925b9
- Add EMeanVarDataType parameter. · f278b2a5
  rocking authored Dec 26, 2022
  
  f278b2a5
26 Dec, 2022 2 commits
- Change of post shuffle thread descriptor · d3f2dbbd
  rocking authored Dec 26, 2022
  
  d3f2dbbd
- Cast only constant · 5490de18
  rocking authored Dec 26, 2022
  
  5490de18
23 Dec, 2022 7 commits
- [What] use ck::math::sqrt instead of __builtin_amdgcn_sqrtf · 7b6c15b9
  rocking authored Dec 23, 2022
```
[Why] __builtin_amdgcn_sqrtf is only support float, double will cause casting
```
  7b6c15b9
- Prevent implicit cast · 4522f7f8
  rocking authored Dec 23, 2022
  
  4522f7f8
- Refine naming · 18e65656
  rocking authored Dec 23, 2022
  
  18e65656
- Refine naming · 44b66c41
  rocking authored Dec 23, 2022
  
  44b66c41
- Pad different size for E and H in layernorm kernel according to different block tile · 14d29856
  rocking authored Dec 23, 2022
  
  14d29856
- Extract divisor out of the loop in reference layernorm · d78877a7
  rocking authored Dec 23, 2022
  
  d78877a7
- Add comment · 62bc2102
  rocking authored Dec 23, 2022
  
  62bc2102
22 Dec, 2022 4 commits
- [What] Force LayernormThreadSliceSize_N = vector size. · 07b9a7b1
  rocking authored Dec 22, 2022
```
[Why] Memory coalesce
```
  07b9a7b1
- Refine var name · 5392333a
  rocking authored Dec 22, 2022
  
  5392333a
- Share the vector size · 441d8973
  rocking authored Dec 22, 2022
  
  441d8973
- Rewrite the 2st kernel, use multiple block along N dimension in layernorm kernel · c83bad81
  rocking authored Dec 22, 2022
  
  c83bad81
21 Dec, 2022 2 commits
- We don't need to pad in N dimension in gemm for mean/var/count. Set NPerTile 1 · 90546fbe
  rocking authored Dec 21, 2022
  
  90546fbe
- Add comment to explan in detail · 119cb7b1
  rocking authored Dec 21, 2022
  
  119cb7b1
20 Dec, 2022 4 commits
- Merge branch 'gemm_layernorm_welford' of... · 19c6cb14
  rocking authored Dec 20, 2022
```
Merge branch 'gemm_layernorm_welford' of https://github.com/ROCmSoftwarePlatform/composable_kernel into gemm_layernorm_welford
```
  19c6cb14
- Refine naming to distinguish naive and welford · 0c0c925e
  rocking authored Dec 20, 2022
  
  0c0c925e
- Add base class for gemm layernorm · d62cd096
  rocking authored Dec 20, 2022
  
  d62cd096
- Merge branch 'develop' into gemm_layernorm_welford · 4d8faca8
  rocking5566 authored Dec 20, 2022
  
  4d8faca8
15 Dec, 2022 8 commits
- Add MNK padding, M = 0 support into grouped_gemm (#539) · 0345963e
  zjing14 authored Dec 15, 2022
```
* add mnk padding, support m=0

* clean code

* clean code
Co-authored-by: Rostyslav Geyyer <46627076+geyyer@users.noreply.github.com>
```
  0345963e
- Merge branch 'develop' into gemm_layernorm_welford · 8572bc7c
  rocking5566 authored Dec 16, 2022
  
  8572bc7c
- Merge branch 'gemm_layernorm_welford' of... · 874a5168
  rocking authored Dec 15, 2022
```
Merge branch 'gemm_layernorm_welford' of https://github.com/ROCmSoftwarePlatform/composable_kernel into gemm_layernorm_welford
```
  874a5168
- Rename the example name · d4d46208
  rocking authored Dec 15, 2022
  
  d4d46208
- disable the attention test that fails on MI100 (#540) · 11151175
  Illia Silin authored Dec 15, 2022
  
  11151175
- Merge branch 'develop' into gemm_layernorm_welford · 904f2a65
  rocking5566 authored Dec 15, 2022
  
  904f2a65
- Add interface GetTypeIdName() and GetTypeIdHashCode() for Device Op (#533) · 10c72ace
  Qianfeng authored Dec 15, 2022
  
  10c72ace
- Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances to enable... · 9a1f2475
  Rostyslav Geyyer authored Dec 14, 2022
```
Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances to enable arbitrary problem size (#535)

* Add padding device_gemm_add_add_fastgelu_xdl_c_shuffle instances

* Add padding device_gemm_add_fastgelu_xdl_c_shuffle instances

* Add gemm_add_fastgelu profiler impl

* Add padding device_gemm_fastgelu_xdl_c_shuffle instances

* Add gemm_fastgelu profiler impl
```
  9a1f2475
14 Dec, 2022 4 commits
- Add a docker hub doc file (#538) · 74744cab
  Rostyslav Geyyer authored Dec 14, 2022
  
  74744cab
- Add pipeline v1/v2 selector · 6eb7610e
  rocking authored Dec 14, 2022
  
  6eb7610e
- Update parameter · 9c66bc6a
  rocking authored Dec 14, 2022
  
  9c66bc6a
- Prevent redundant IO · bb7294ae
  rocking authored Dec 14, 2022
  
  bb7294ae