Initial opimization/Performance improvements for gemm#87

Open

hroaten wants to merge 10 commits into

mainfrom

Commits on Apr 15, 2026

minmaxmean
Holden Roaten
committed
Sort of optimized for handling scalars equalling 0 or 1
Holden Roaten
committed

Commits on Apr 16, 2026

gemm quick returns
Holden Roaten
committed
implemented basic memory coalescing
Holden Roaten
committed

Commits on Apr 17, 2026

Implemented shared-memory cache blocking
Holden Roaten
committed
zero_device kernel for zeroing array
Holden Roaten
committed

Commits on Apr 20, 2026

2x naive perforamnce inmprovement via shared memory cache blocking
Holden Roaten
committed
actually fixed merge conflicts
Holden Roaten
committed
reverted unintended changes
Holden Roaten
committed
Merge branch 'main' into hroaten/gemm
tdehoff
authored