Skip to content

qwen3_5 cuda: 2, 3, 5 and 6-bit tiles of their own width and a prompt GEMM for them (27B oQ2 on 16 GB) - #859

Merged
ashhart merged 36 commits into
ashhart:mainfrom
BobClawblaw:cuda-native-lowbit
Oct 18, 2026
Merged

ashhart merged 36 commits into
ashhart:mainfrom
BobClawblaw:cuda-native-lowbit

qwen3_5 cuda: the load staging buffer freed after load

62a7b55
Select commit
Loading
Failed to load commit list.
Sign in for the full log view

Annotations

1 warning and 1 notice
zig
succeeded Oct 16, 2026 in 5m 35s