Commits · 39821a903db74d88893177279b4e6c9e6dbde2b8 · gaoqiong / composable_kernel

29 Jan, 2021 1 commit
- refactor, clean up · 39821a90
  Chao Liu authored Jan 29, 2021
  
  39821a90
28 Jan, 2021 2 commits
- fix control flow issue for padding case · 01e94729
  Chao Liu authored Jan 28, 2021
  
  01e94729
- add vector load and access order to threadwise copy v1r3 · 68ea43b1
  Chao Liu authored Jan 27, 2021
  
  68ea43b1
21 Jan, 2021 1 commit
- added vector load, but ISA is much worse · 5fdccfce
  Chao Liu authored Jan 20, 2021
  
  5fdccfce
19 Jan, 2021 1 commit
- adding vector load · 44ddcdcb
  Chao Liu authored Jan 18, 2021
  
  44ddcdcb
17 Jan, 2021 1 commit
- improving index calculation · 77c81617
  Chao Liu authored Jan 16, 2021
  
  77c81617
11 Jan, 2021 1 commit
- add hack for merge transformation · f2f35201
  Chao Liu authored Jan 11, 2021
  
  f2f35201
08 Jan, 2021 1 commit
- modify gridwise dynamic gemm looping · fa479ce4
  Chao Liu authored Jan 07, 2021
  
  fa479ce4
03 Dec, 2020 1 commit
- use StaticallyIndexedArray for buffer in threadwise copy, in order to get rid of alloca in IR · 760a234f
  Chao Liu authored Dec 03, 2020
  
  760a234f
26 Nov, 2020 1 commit
- add dynamic forward v4r4, but see lots of scrach mem · be4e3133
  Chao Liu authored Nov 25, 2020
  
  be4e3133
20 Nov, 2020 1 commit
- fix scratch memory issue · 8e5085fa
  Chao Liu authored Nov 19, 2020
  
  8e5085fa
08 Nov, 2020 1 commit
- tweak bwd-data-v4r1 · b50fa980
  Chao Liu authored Nov 07, 2020
  
  b50fa980
26 Oct, 2020 1 commit
- debugging scatch mem issue · 92661018
  Chao Liu authored Oct 25, 2020
  
  92661018
15 Oct, 2020 1 commit
- update buffer addressing · 5af722b4
  Chao Liu authored Oct 14, 2020
  
  5af722b4
08 Oct, 2020 1 commit
- adding dynamic copy · 711701f3
  Chao Liu authored Oct 08, 2020
  
  711701f3
05 Oct, 2020 1 commit
- fix · 176d9fc0
  Chao Liu authored Oct 05, 2020
  
  176d9fc0
02 Oct, 2020 1 commit
- use statically index array for all existing kernels · 5e88414a
  Chao Liu authored Oct 02, 2020
  
  5e88414a
01 Oct, 2020 4 commits
- fixed a bug; tested more transformation, no scrach mem · a578ff93
  Chao Liu authored Oct 01, 2020
  
  a578ff93
- refactor most loop to static · c15a3c09
  Chao Liu authored Oct 01, 2020
  
  c15a3c09
- refactor · 9ae20fc2
  Chao Liu authored Oct 01, 2020
  
  9ae20fc2
- refactored dynamically indexed array works now · 34cbbb48
  Chao Liu authored Sep 30, 2020
  
  34cbbb48
30 Sep, 2020 2 commits
- testing · 5a2498d1
  Chao Liu authored Sep 30, 2020
  
  5a2498d1
- refactoring array, tuple · e7f633c5
  Chao Liu authored Sep 30, 2020
  
  e7f633c5
29 Sep, 2020 1 commit
- refactoring array · 6cd94d98
  Chao Liu authored Sep 29, 2020
  
  6cd94d98
26 Sep, 2020 3 commits
- refactoring array · 9535f806
  Chao Liu authored Sep 26, 2020
  
  9535f806
- refactoring array type · c98cbea0
  Chao Liu authored Sep 26, 2020
  
  c98cbea0
- refactor · 674c405f
  Chao Liu authored Sep 25, 2020
  
  674c405f
24 Sep, 2020 1 commit
- refactor · fc7a6c85
  Chao Liu authored Sep 23, 2020
  
  fc7a6c85
18 Sep, 2020 1 commit
- trying to solve scratch mem issue · f3099df8
  Chao Liu authored Sep 18, 2020
  
  f3099df8
13 Sep, 2020 1 commit
- debugging scratcg mem · 0413a020
  Chao Liu authored Sep 13, 2020
  
  0413a020
11 Sep, 2020 2 commits
- clean up · 0a944e8f
  Chao Liu authored Sep 11, 2020
  
  0a944e8f
- adding dynamic tensor descriptor · 4e78d2fc
  Chao Liu authored Sep 11, 2020
  
  4e78d2fc
18 Aug, 2020 1 commit
- prototype dynamic tensor descriptor · 5514dc0b
  Chao Liu authored Aug 18, 2020
  
  5514dc0b
29 Jul, 2020 1 commit
- prototype dynamic descriptor · d3405258
  Chao Liu authored Jul 29, 2020
  
  d3405258
02 Jul, 2020 1 commit
- experiment dummy static transform · bd4c668e
  Chao Liu authored Jul 02, 2020
  
  bd4c668e
24 Jun, 2020 1 commit

Chao Liu authored Jun 23, 2020



* tuning para,

* testing on v100

* add fp16

* remove deprecated tensor descriptor

* sync with miopen

* update build script
Co-authored-by: Jing Zhang <jizhan@amd.com>

5c7cec11

27 Jan, 2020 1 commit
- Update for recent MIOpen integration (#11) · 3406a114
  Chao Liu authored Jan 27, 2020
```
* update for MIOpen integration
```
  3406a114
20 Jan, 2020 1 commit

Added bwd data v3r1 v4r1, tweaking v1 (#10) · c5da0377

Chao Liu authored Jan 20, 2020

* Added bwd data v3r1: breaking down compute into a series of load balanced GEMM, and launch in a single kernel
* Added bwd data v4r1: like v3r1, but launch GEMMs in multiple kernels
* Tweaked v1r1  and v1r2 (atomic) on AMD GPU

c5da0377

05 Dec, 2019 1 commit
- update implicit GEMM forward v4r4 to use gridwise gemm (#9) · e2b4c5b4
  Chao Liu authored Dec 05, 2019
```
* updated fwd v4r4 to use gridwise gemm
* updated gridwise gemm api calls in bwd-data v1r1 and v2r1
```
  e2b4c5b4
03 Dec, 2019 1 commit

backward data (#7) · 8f5f6496

Chao Liu authored Dec 03, 2019

* enabled atomic add in tensor copy
* added gridwise GEMM
* added backward data conv using GEMM + atomic
* added backward data conv using GEMM, no atomic

8f5f6496