[LoRA] Update LoRA expand kernel block_n calculation (#32621)

Signed-off-by: Xin Yang <xyangx@amazon.com>

[LoRA] Update LoRA expand kernel block_n calculation (#32621)
Signed-off-by: Xin Yang <xyangx@amazon.com>
c870eb9e · Xin Yang · GitHub · 6af03f23 · c870eb9e
Unverified Commit c870eb9e authored Feb 23, 2026 by Xin Yang Committed by GitHub Feb 23, 2026
Show whitespace changes
Inline Side-by-side

Showing with 1 addition and 1 deletion

vllm/lora/ops/triton_ops/utils.py vllm/lora/ops/triton_ops/utils.py +1 -1

No files found.
--- a/vllm/lora/ops/triton_ops/utils.py
+++ b/vllm/lora/ops/triton_ops/utils.py
@@ -251,7 +251,7 @@ def get_lora_op_configs(
    else:
        default = {
            "block_m": 64,
-            "block_n": max(64, next_power_of_2(128 // num_slices)),
+            "block_n": 64 if num_slices > 1 else 128,
            "block_k": 16,
            "num_warps": 4,
            "num_ctas": 1,