Skip to content

test: skip NVFP4-KV decode tests on SM107 #4257

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

Merged
jimmyzho merged 1 commit into from
Jul 30, 2026
+8 −0
Merged

test: skip NVFP4-KV decode tests on SM107 #4257

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
8 changes: 8 additions & 0 deletions tests/attention/test_batch_decode_kernels.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,13 @@ def skip_if_nvfp4_large_head_decode_unsupported(head_dim: int):
)


def skip_if_nvfp4_kv_unsupported():
# Mirrors the guard in flashinfer.decode/prefill: no NVFP4-KV decode
# kernels exist for SM107.
if get_compute_capability(torch.device("cuda:0")) == (10, 7):
pytest.skip("KV Cache NVFP4 is not supported on SM107")


@pytest.fixture(
autouse=not has_flashinfer_jit_cache(),
scope="module",
Expand Down Expand Up @@ -734,6 +741,7 @@ def test_batch_decode_with_paged_kv_cache_nvfp4(
Reference is computed by dequantizing the packed KV back to q_dtype and running
single_decode_with_kv_cache per batch item.
"""
skip_if_nvfp4_kv_unsupported()
kv_layout = "NHD"
torch.manual_seed(42)

Expand Down
Loading