add support for 64 block size on 32 warp size supported amd gpus - #1748

Merged

matthewdouglas merged 14 commits into

bitsandbytes-foundation:mainfrom

electron271:main

Nov 13, 2025

Commits on Sep 6, 2025

add support for 64 block size on 32 warp size supported amd gpus
electron271
committed
d607127
uncomment 64 block size support in csrc
electron271
committed
f7b4430
only enable 64 block size support on architectures with 32 warp size
electron271
committed
6e2e4d2

Commits on Oct 2, 2025

use BNB_WARP_SIZE instead of warpSize in ops.hip
electron271
committed
8c24b4d

Commits on Oct 20, 2025

Reuse BNB_WARP_SIZE macro
sstamenk
committed
7dd7b88
Merge pull request #1 from sstamenk/reuse_bnb_warp_size_macro
electron271
authored
978eccf
Merge branch 'main' into main
electron271
authored
1c94701

Commits on Oct 24, 2025

Remove unused WARP_SIZE definitions
electron271
committed
e9f0af3

Commits on Oct 29, 2025

Merge branch 'main' into main
matthewdouglas
authored
ca63206

Commits on Nov 5, 2025

remove unused import
electron271
committed
39861c2
Merge branch 'main' into main
electron271
authored
6f6c5e1

Commits on Nov 13, 2025

Apply suggestion from @matthewdouglas
matthewdouglas
authored
7f41f0e
Apply suggestion from @matthewdouglas
matthewdouglas
authored
22685ac
Apply suggestion from @matthewdouglas
matthewdouglas
authored
f0a7c13