Commits · 51884fc21412b1800bb85b28c0f5a0b651d23cef · yangql / composable_kernel-1

22 Sep, 2019 1 commit
- WIP: explicitly separate offset component into compile-time, block-invariant... · 51884fc2
  Chao Liu authored Sep 21, 2019
```
WIP: explicitly separate offset component into compile-time, block-invariant and per-thread components
```
  51884fc2
20 Sep, 2019 1 commit
- refactor · bf7e7d62
  Chao Liu authored Sep 19, 2019
  
  bf7e7d62
19 Sep, 2019 1 commit
- use buffer_load buffer_store intrinsic · b6e1c52a
  Chao Liu authored Sep 19, 2019
  
  b6e1c52a
18 Sep, 2019 5 commits
- refactor · 94bb1b48
  Chao Liu authored Sep 18, 2019
  
  94bb1b48
- add global_load and buffer_load inline asm · 86cc678f
  Chao Liu authored Sep 18, 2019
  
  86cc678f
- experimenting global and buffer load/store · 5b7a18c5
  Chao Liu authored Sep 18, 2019
  
  5b7a18c5
- experimenting global and buffer load/store · c7a6545e
  Chao Liu authored Sep 18, 2019
  
  c7a6545e
- experimenting global and buffer load/store · 9f46cdf5
  Chao Liu authored Sep 18, 2019
  
  9f46cdf5
17 Sep, 2019 3 commits
- enable hip compiler flag: -amdgpu-enable-global-sgpr-addr · f58bf384
  Chao Liu authored Sep 17, 2019
  
  f58bf384
- bug fix · 126cae0c
  Chao Liu authored Sep 17, 2019
  
  126cae0c
- refactor · f7be86b9
  Chao Liu authored Sep 16, 2019
  
  f7be86b9
16 Sep, 2019 1 commit
- bug fix · d7079939
  Chao Liu authored Sep 15, 2019
  
  d7079939
15 Sep, 2019 5 commits
- amd build · 69fea593
  Chao Liu authored Sep 15, 2019
  
  69fea593
- add lds doble buffer to nchw padded v4r1 and v4r4 · 940949d9
  Chao Liu authored Sep 15, 2019
  
  940949d9
- add lds doble buffer to nchw padded v4r1 and v4r4 · bf975428
  Chao Liu authored Sep 15, 2019
  
  bf975428
- initial implementation for nchw v4r4 padding · 2c93b305
  Chao Liu authored Sep 15, 2019
  
  2c93b305
- clean up · 53094f7f
  Chao Liu authored Sep 15, 2019
  
  53094f7f
14 Sep, 2019 1 commit
- initial padding support for nchw · d4878d99
  Chao Liu authored Sep 13, 2019
  
  d4878d99
12 Sep, 2019 2 commits
- clean up · bd7a2300
  Chao Liu authored Sep 12, 2019
  
  bd7a2300
- padding for chwn is functional · 1f705244
  Chao Liu authored Sep 12, 2019
  
  1f705244
11 Sep, 2019 1 commit
- enabling padding for chwn format · 724e984b
  Chao Liu authored Sep 11, 2019
  
  724e984b
10 Sep, 2019 1 commit
- adding merge transform · ca42e910
  Chao Liu authored Sep 10, 2019
  
  ca42e910
09 Sep, 2019 1 commit
- more utility code · 7a7fe160
  Chao Liu authored Sep 09, 2019
  
  7a7fe160
06 Sep, 2019 1 commit
- added tuple · 625838de
  Chao Liu authored Sep 06, 2019
  
  625838de
23 Aug, 2019 1 commit
- clean · cb6475c7
  Chao Liu authored Aug 23, 2019
  
  cb6475c7
20 Aug, 2019 1 commit
- adding tensor_view · 238d58c2
  Chao Liu authored Aug 20, 2019
  
  238d58c2
15 Aug, 2019 1 commit
- clean up · 0979fb4a
  Chao Liu authored Aug 15, 2019
  
  0979fb4a
14 Aug, 2019 1 commit
- adding padding to implicit gemm v1r3 · 4fb81e00
  Chao Liu authored Aug 14, 2019
  
  4fb81e00
13 Aug, 2019 3 commits
- clean up · 740149fc
  Chao Liu authored Aug 13, 2019
  
  740149fc
- add back some code · 40836ab9
  Chao Liu authored Aug 13, 2019
  
  40836ab9
- clean up · 8bdaba51
  Chao Liu authored Aug 13, 2019
  
  8bdaba51
12 Aug, 2019 1 commit
- clean up · fab2f10a
  Chao Liu authored Aug 12, 2019
  
  fab2f10a
10 Aug, 2019 1 commit
- cleaning up · 1c4ef23c
  Chao Liu authored Aug 09, 2019
  
  1c4ef23c
08 Aug, 2019 4 commits
- tweak on amd · 4908fe3f
  Chao Liu authored Aug 08, 2019
  
  4908fe3f
- added ThreadwiseGenericTensorSliceCopy_v2r1 · a9b2b1dc
  Chao Liu authored Aug 08, 2019
  
  a9b2b1dc
- clean up · 701b7341
  Chao Liu authored Aug 07, 2019
  
  701b7341
- use ford/for instead of static_ford/static_for in threadwise copy, somehow... · bc9ea646
  Chao Liu authored Aug 07, 2019
```
use ford/for instead of static_ford/static_for in threadwise copy, somehow register spill is greatly reduced on AMD
```
  bc9ea646
07 Aug, 2019 3 commits
- bug fix in ford, forgot to reorder lengths · 5636576f
  Chao Liu authored Aug 07, 2019
  
  5636576f
- adding ThreadwiseGenericTensorSliceCopy_v1r2 · 9d99a580
  Chao Liu authored Aug 07, 2019
  
  9d99a580
- reworked ThreadwiseGenericTensorSliceCopy_v1 · 1b3c2e40
  Chao Liu authored Aug 07, 2019
  
  1b3c2e40