Weixing Zhang 34b89a1173 [OPTIMIZER] Tweak warpsPerCTA based on the shape of MMA output (#2581) ...

In current implementation, warpsPerCTA is always set to [numWarps, 1]
for 2 tt.dot fusion scenario. But, it is not optimal for cases such that
tt.dot doesn't have enough parallelism on row dimension but on column
dimension.

2023-11-03 16:40:03 -04:00

..

examples

Reformat Python code with yapf. (#2589)

2023-11-02 20:44:17 -07:00

src

[OPTIMIZER] Thread local reduction optimization (#2542)

2023-10-31 16:13:36 -07:00

test

[OPTIMIZER] Tweak warpsPerCTA based on the shape of MMA output (#2581)

2023-11-03 16:40:03 -04:00

triton

[OPTIMIZATION] Enable pipelining for bwd flash attention (#2590)

2023-11-03 11:46:51 -07:00

tutorials

[OPTIMIZATION] Enable pipelining for bwd flash attention (#2590)

2023-11-03 11:46:51 -07:00

MANIFEST.in

[TOOLS] improved ahead-of-time compiler (#1805)

2023-06-21 01:02:58 -07:00

pyproject.toml

Enable ruff linter instead of flake8 (#2574)

2023-10-31 21:28:24 +00:00

setup.py

Reformat Python code with yapf. (#2589)

2023-11-02 20:44:17 -07:00