Amir Ghamarian
5b591af58e
Add block_size=32 support and fix int32 overflow in unified attention
...
- Add BlockSize_ template parameter to all kernel trait structs,
enabling block_size=32 instances for d64/GQA-8 (medium/small/tiny tiers).
- 12 new instance files for block_size=32 (bf16/fp16 x mask/nmask x 3 tiers).
- Dispatch branches on page_blk_size to select bs32 vs bs64 instances.
- Fix potential int32 overflow in KV cache tensor view by using long_index_t
for sizes and strides (overflow threshold was ~66K blocks for d64/GQA-8).
Made-with: Cursor
2026-03-30 14:16:17 +00:00
..
2026-01-15 07:19:31 -08:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-12-18 07:59:45 +01:00
2025-11-28 13:49:54 -08:00
2025-10-16 03:10:57 -07:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2026-01-15 16:43:02 +01:00
2026-01-13 07:14:23 +01:00
2026-01-07 16:30:57 +01:00
2025-12-18 07:59:45 +01:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-12-18 07:59:45 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2026-01-17 08:30:27 +01:00
2025-12-15 07:16:01 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2026-01-17 08:30:27 +01:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2025-12-30 16:25:08 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2026-01-07 16:30:57 +01:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2025-12-18 13:12:15 -07:00
2025-11-28 13:49:54 -08:00
2025-11-28 13:49:54 -08:00
2026-01-15 16:43:02 +01:00
2025-11-28 13:49:54 -08:00
2026-01-07 16:30:57 +01:00
2026-01-15 16:43:02 +01:00
2026-01-15 16:43:02 +01:00
2026-03-30 14:16:17 +00:00
2026-01-14 07:31:45 -08:00
2024-12-04 00:46:47 +01:00