Batched NNLS - #8402

Draft

achirkin wants to merge 13 commits into

NVIDIA:mainfrom

achirkin:fea-nnls

Commits on Jul 17, 2026

wip: many draft implementations
achirkin
committed
8655a2c
Use raft block reductions
achirkin
committed
974dfac
Incremental cholesky updates and G in global memory / L2
achirkin
committed
cec0066
Flexible kernel dispatch
achirkin
committed
2194eca
Warp-level Cholesky
achirkin
committed
a8f4edf
More incremental Cholesky solve
achirkin
committed
fbbf4c1
Keep the decomposition matrices in global memory
achirkin
committed
6632856

Commits on Jul 20, 2026

A few more tweaks to parallelize cholesky
achirkin
committed
4b8c4dc
Cleanup
achirkin
committed
9a36d0a
Merge branch 'main' into fea-nnls
achirkin
committed
8647335
Remove cumlarray
achirkin
committed
3d36f1d

Commits on Jul 22, 2026

Merge branch 'main' into fea-nnls
achirkin
authored
e19e97c
Fix style
achirkin
committed
d1e2da2