Commits · ee77b931b91476a41e4b1c19cf76281444044908 · gaoqiong / flash-attention

11 Sep, 2023 1 commit
- Swap seqlen_q and nheads for MQA to speed it up (h/t Daniel Haziza) · ee77b931
  Tri Dao authored Sep 10, 2023
  
  ee77b931
04 Sep, 2023 5 commits
- Implement flash_attn_with_kvcache · 37c6e054
  Tri Dao authored Sep 04, 2023
  
  37c6e054
- Remove constexpr in launch template to fix CI compilation · 6a89b2f1
  Tri Dao authored Sep 03, 2023
  
  6a89b2f1
- Try switching back to Cutlass 3.2.0 · 97ba7a62
  Tri Dao authored Sep 03, 2023
  
  97ba7a62
- Bump to v2.1.2 · 1dc1b6c8
  Tri Dao authored Sep 03, 2023
  
  1dc1b6c8
- Remove unused sdPsum in dot_do_o function · 5953c4f5
  Tri Dao authored Sep 03, 2023
  
  5953c4f5
03 Sep, 2023 1 commit
- Fix splitKV combine function when local LSEs are all -inf · 26d7d92f
  Tri Dao authored Sep 03, 2023
  
  26d7d92f
01 Sep, 2023 2 commits
- Remove commented out code in bwd (#512) · 37e32feb
  Sophia Wisdom authored Sep 01, 2023
```
* Remove lots of comments

* Remove unused traits
```
  37e32feb
- Remove old code in utils.h (#511) · dd8a7549
  Sophia Wisdom authored Sep 01, 2023
  
  dd8a7549
30 Aug, 2023 2 commits
- bump cutlass submodule (#504) · 866a9d33
  Aman Gupta Karmani authored Aug 30, 2023
  
  866a9d33
- Fix typo with lse_max == -INFINITY · 31920dda
  Tri Dao authored Aug 29, 2023
  
  31920dda
29 Aug, 2023 1 commit
- Implement splitKV attention · b1fbbd83
  Tri Dao authored Aug 29, 2023
  
  b1fbbd83
28 Aug, 2023 3 commits
- Use generate_kernels.py script from Driss Guessous · 7a983df7
  Tri Dao authored Aug 28, 2023
  
  7a983df7
- [ft_attention] Fix for seqlen=8136 (#488) · c3f2a632
  dan_the_3rd authored Aug 28, 2023
```
When seqlen=8136, `smem_sz = 48840`, and apparently starting the kernel returns an `invalid argument` CUDA error.

`48840 < 48 * 1024` but apparently it's still above the limit somehow..?
Tested on A100
```
  c3f2a632
- Update Cutlass to v3.2.0 · 757058d4
  Tri Dao authored Aug 27, 2023
  
  757058d4
25 Aug, 2023 1 commit
- Change causal mask to be aligned to bottom-right instead of top-left · 9e5e8bc9
  Tri Dao authored Aug 21, 2023
  
  9e5e8bc9
24 Aug, 2023 1 commit
- Support flash attention 2 with causal masking when KV's seq length is longer... · e07aa036
  BoxiangW authored Aug 24, 2023
```
Support flash attention 2 with causal masking when KV's seq length is longer than Q's seq length. (#436)
```
  e07aa036
17 Aug, 2023 1 commit
- [FusedDense] Run black on fused_dense.py · bcfa7c97
  Tri Dao authored Aug 16, 2023
  
  bcfa7c97
16 Aug, 2023 1 commit
- Fix Bwd NaN for varlen when seqlen_q >> seqlen_k and causal · c65b5106
  Tri Dao authored Aug 16, 2023
  
  c65b5106
13 Aug, 2023 2 commits
- Prepare for Cutlass 3.2 · dbd79237
  Tri Dao authored Aug 13, 2023
  
  dbd79237
- Update to Cutlass 3.1 · 3524e13c
  Tri Dao authored Aug 13, 2023
  
  3524e13c
01 Aug, 2023 2 commits
- Fix race condition in bwd (overwriting sK) · 1c41d2b0
  Tri Dao authored Aug 01, 2023
  
  1c41d2b0
- Fix masking of bwd when seqlen is not divisible by 128 · a4f148b6
  Tri Dao authored Jul 31, 2023
  
  a4f148b6
27 Jul, 2023 1 commit

Kirthi Shankar Sivamani authored Jul 27, 2023



* Add RNG state to kernel launch params
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Save seed and offset for backward
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Single thread write to global mem
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* compute_dq_dk_dv_1colblock get seed and offset from launch params
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* compute_dq_dk_dv_1rowblock get seed and offset from launch params
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Change forward c++ APIs to save RNG state for backward
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Change backward c++ APIs to set RNG state for bprop launcher
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Bug fixes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Python side API changes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Bug fix; only save seeds instead of full offset
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Account for 3D grid size
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

a03f6f8e

23 Jul, 2023 2 commits
- Fix random state for dropout_layer_norm (#315) · 767b71cc
  Joel Lamy-Poirier authored Jul 23, 2023
  
  767b71cc
- [FT] Implement MQA/GQA · a157cc8c
  Tri Dao authored Jul 22, 2023
  
  a157cc8c
21 Jul, 2023 1 commit
- Fix using dO stride for O, which can cause memory error in bwd · 9ee0ff1d
  Tri Dao authored Jul 20, 2023
  
  9ee0ff1d
19 Jul, 2023 2 commits
- [LayerNorm] Fix typo in ln_api.cpp · dfc60f6b
  Ikko Eltociear Ashimine authored Jul 20, 2023
```
unintialized -> uninitialized
```
  dfc60f6b
- Fix compile error on MSVC · 538d570c
  danthe3rd authored Jul 19, 2023
```
See also: https://stackoverflow.com/questions/55136414/constexpr-variable-captured-inside-lambda-loses-its-constexpr-ness
```
  538d570c
17 Jul, 2023 1 commit
- FlashAttention-2 release · 4f285b35
  Tri Dao authored Jul 17, 2023
  
  4f285b35
06 Jul, 2023 1 commit
- [FT] rotary_cos/sin should have batch_size dimension · 2800efc7
  Tri Dao authored Jul 06, 2023
  
  2800efc7
03 Jul, 2023 1 commit
- [FT] rotary_cos/sin should have shape (dim) instead of (seqlen, dim) · 3a9bfd07
  Tri Dao authored Jul 03, 2023
  
  3a9bfd07
02 Jul, 2023 1 commit
- [Rotary] Make sure frequency calculation is in fp32 · 62e98144
  Tri Dao authored Jul 02, 2023
  
  62e98144
30 May, 2023 2 commits
- [FusedDense] Allocate lt_workspace on input device · 27f8f890
  Tri Dao authored May 30, 2023
  
  27f8f890
- [Gen] Add rotary base as an argument to FT attention kernel · 48bc6eac
  Tri Dao authored May 30, 2023
  
  48bc6eac
26 Apr, 2023 1 commit
- [Docs] Clearer error message for bwd d > 64, bump to v1.0.4 · ad113948
  Tri Dao authored Apr 26, 2023
  
  ad113948
21 Apr, 2023 1 commit
- [Gen] Fix FT kernel smem size, CG when batch size changed · 311d6606
  Tri Dao authored Apr 20, 2023
  
  311d6606
15 Apr, 2023 1 commit
- only 1 thread writes to global mem in fprop · 45567a25
  Kirthi Shankar Sivamani authored Apr 15, 2023
```
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
  45567a25
13 Apr, 2023 1 commit
- Handle FlashAttnQKVPackedSplitFunc by making rng_state optional in backward · 7d25a4ec
  Kirthi Shankar Sivamani authored Apr 13, 2023
```
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
  7d25a4ec
12 Apr, 2023 1 commit
- Support CUDA graph capture · 31018c5f
  Kirthi Shankar Sivamani authored Apr 12, 2023
```
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
  31018c5f