Vinayak Gokhale e0a4d97569 Mask vs pad for non power of 2 sequence lengths ...

Padding results in memory allocation which is slower.
Masking results in better performance.

2023-12-04 10:11:41 -06:00

..

examples

[FORMAT] autopep8 now uses max-line-length=88 (#1410)

2023-03-25 15:46:50 -07:00

perf-kernels

Mask vs pad for non power of 2 sequence lengths

2023-12-04 10:11:41 -06:00

src

Merge commit '721897fcc4f942aa97d2e9ba3787a5e213758177' into ifu-231108

2023-11-08 18:51:23 +00:00

test

use hw for fp8 type conversion (#386)

2023-11-24 10:26:40 -06:00

triton

use hw for fp8 type conversion (#386)

2023-11-24 10:26:40 -06:00

tutorials

ROCM IFU: Resolve merge conflicts in tutorial 06

2023-11-17 01:28:40 +00:00

MANIFEST.in

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

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

pyproject.toml

[BUILD] use ninja (#2318)

2023-09-18 15:30:06 -05:00

setup.py

Merge commit '721897fcc4f942aa97d2e9ba3787a5e213758177' into ifu-231108

2023-11-08 18:51:23 +00:00