Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix max_seqlens_q/k initialization for Navi GPUs #310

Merged
merged 1 commit into from
Dec 9, 2024
Merged

Fix max_seqlens_q/k initialization for Navi GPUs #310

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 4 additions & 5 deletions vllm/attention/ops/triton_flash_attention.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -912,9 +912,8 @@ def check_and_convert(t, scale):
p_descale = 1.0 / p_scale
o_descale = 1.0 / o_scale

if is_navi():
max_seqlens_q = 0
max_seqlens_k = 0
arg_max_seqlens_q = 0 if is_navi() else max_seqlens_q
arg_max_seqlens_k = 0 if is_navi() else max_seqlens_k

attn_fwd[grid](
q,
Expand Down Expand Up @@ -944,8 +943,8 @@ def check_and_convert(t, scale):
HQ=nheads_q,
HK=nheads_k,
ACTUAL_BLOCK_DMODEL=head_size,
MAX_SEQLENS_Q=max_seqlens_q,
MAX_SEQLENS_K=max_seqlens_k,
MAX_SEQLENS_Q=arg_max_seqlens_q,
MAX_SEQLENS_K=arg_max_seqlens_k,
IS_CAUSAL=causal,
VARLEN=True,
BLOCK_DMODEL=padded_d_model,
Expand Down
Loading