【Fix bug] w4afp8 的nblock固定为256，并且fa3的append attn 增加mask参数 (#3771) (#3835)

* fix w4afp8 * 增加集中式配置 * codestyle * fix fa3 append attn
2025-10-05 08:37:06 +08:00 · 2025-09-03 19:36:45 +08:00
parent 87ef0f5d30
commit 9213a58a06
3 changed files with 4 additions and 5 deletions
--- a/custom_ops/utils/auto_gen_w4afp8_gemm_kernel.py
+++ b/custom_ops/utils/auto_gen_w4afp8_gemm_kernel.py
@@ -88,6 +88,8 @@ gemm_case = [
    [8192, 3584, 8, 2048],  # eb45T ffn1
    [7168, 8192, 8, 0],  # eb45T ffn2
    [7168, 8192, 8, 2048],  # eb45T ffn2
+    [1792, 8192, 64, 0],  # eb45t ffn1
+    [8192, 896, 64, 0],  # eb45t ffn2
 ]

 dtype = ["BF16"]