Gabriel Wu 2d8c4f22d5 fix: windows compat ...

Signed-off-by: Gabriel Wu <13583761+lucifer1004@users.noreply.github.com>

2025-04-23 14:47:15 +08:00

..

include/deep_gemm

refactor: compile to .cubin and add NVRTC option

2025-04-22 10:17:52 +00:00

jit

fix: windows compat

2025-04-23 14:47:15 +08:00

jit_kernels

[wip] refactor: compile to .cubin

2025-04-22 08:08:40 +00:00

__init__.py

fix typo

2025-02-26 18:37:22 +08:00

utils.py

Correctly flush L2, as reconstructing the tensors on every iteration effectively put them in the L2, and gave the GPU enough idle time to avoid thermal throttling in a potentially unrealistic way.

2025-03-15 20:46:24 +00:00