Commits · 562e1e27673bea4d9ce6793d418c7788138e49ed · yangql / composable_kernel-1

04 Nov, 2019 1 commit
- MIOpen integration: recent bug fixes from MIOpen (#5) · 562e1e27
  Chao Liu authored Nov 04, 2019
  
  562e1e27
11 Oct, 2019 1 commit
- Refactor for MIOpen integration (#4) · 52c3fe05
  Chao Liu authored Oct 11, 2019
```
Refactor, so can bring multi-index transformation and padding support into MIOpen
```
  52c3fe05
27 Sep, 2019 4 commits
- tweaking · 012d3a07
  Chao Liu authored Sep 27, 2019
  
  012d3a07
- debugging · ebe38f3d
  Chao Liu authored Sep 27, 2019
  
  ebe38f3d
- remove dead code · 9b280cc5
  Chao Liu authored Sep 27, 2019
  
  9b280cc5
- nvidia build · 98a2cfcc
  Chao Liu authored Sep 27, 2019
  
  98a2cfcc
26 Sep, 2019 3 commits
- removing dependency on old tensor descriptor · 51a9fa1d
  Chao Liu authored Sep 26, 2019
  
  51a9fa1d
- added type conversion in threadwise and blockwise copy · b3d4595f
  Chao Liu authored Sep 25, 2019
  
  b3d4595f
- removing old implementation of tensor descriptor · 39d92e7d
  Chao Liu authored Sep 25, 2019
  
  39d92e7d
25 Sep, 2019 2 commits
- added GetLinearDimensionMask · e1ae8f18
  Chao Liu authored Sep 25, 2019
  
  e1ae8f18
- adding GetLinearDimensionMask() · 4f4aba48
  Chao Liu authored Sep 24, 2019
  
  4f4aba48
24 Sep, 2019 1 commit
- refactor · 545d9305
  Chao Liu authored Sep 24, 2019
  
  545d9305
22 Sep, 2019 3 commits
- nvidia build · 37f4e2b6
  Chao Liu authored Sep 22, 2019
  
  37f4e2b6
- done: explicitly separate offset component into compile-time, block-invariant... · 6c2c50b0
  Chao Liu authored Sep 22, 2019
```
done: explicitly separate offset component into compile-time, block-invariant and per-thread components. Experimenting
```
  6c2c50b0
- 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
21 Sep, 2019 3 commits
- refactor · 740da00a
  Chao Liu authored Sep 20, 2019
  
  740da00a
- nvidia build · 184c6e7d
  Chao Liu authored Sep 20, 2019
  
  184c6e7d
- adding logic to judge linear dimension · f00c1381
  Chao Liu authored Sep 20, 2019
  
  f00c1381
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 4 commits
- 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
15 Sep, 2019 1 commit
- amd build · 69fea593
  Chao Liu authored Sep 15, 2019
  
  69fea593
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
05 Sep, 2019 1 commit
- adding dimension tranformation · 0c05f427
  Chao Liu authored Sep 05, 2019
  
  0c05f427
02 Sep, 2019 1 commit
- adding dimension transformation · bd44e639
  Chao Liu authored Sep 02, 2019
  
  bd44e639
20 Aug, 2019 1 commit
- adding tensor_view · 238d58c2
  Chao Liu authored Aug 20, 2019
  
  238d58c2
13 Aug, 2019 1 commit
- clean up · 740149fc
  Chao Liu authored Aug 13, 2019
  
  740149fc
08 Aug, 2019 2 commits
- added ThreadwiseGenericTensorSliceCopy_v2r1 · a9b2b1dc
  Chao Liu authored Aug 08, 2019
  
  a9b2b1dc
- 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
- reworked ThreadwiseGenericTensorSliceCopy_v1 · 1b3c2e40
  Chao Liu authored Aug 07, 2019
  
  1b3c2e40
- add looping Orders into ford and static_ford · 41cdde99
  Chao Liu authored Aug 06, 2019
  
  41cdde99
06 Aug, 2019 2 commits
- added ReorderGiveOld2New() in Sequence and ConstantTensorDescriptor · 0271338e
  Chao Liu authored Aug 06, 2019
  
  0271338e
- reimplement threadwise copy · fdcfae3a
  Chao Liu authored Aug 06, 2019
  
  fdcfae3a