Skip to content

[Triton/Gluon] Skip MXFP4 logits tests on unsupported GPUs #5958

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
azaidy merged 1 commit into from
Oct 1, 2026
Merged

[Triton/Gluon] Skip MXFP4 logits tests on unsupported GPUs #5958

Show file tree
Hide file tree
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
6 changes: 6 additions & 0 deletions op_tests/triton_tests/attention/test_pa_mqa_logits_mxfp4.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,8 +11,14 @@
mfma_nonk_dim,
paged_mxfp4_mqa_logits,
)
from aiter.ops.triton.utils._triton import arch_info
from op_tests.triton_tests.utils.pa_mqa_logits_mxfp4_ref import preshuffle_cache

pytestmark = pytest.mark.skipif(
arch_info.get_arch() != "gfx950",
reason="paged MXFP4 MQA logits is gfx950 only",
)

SCALE_GROUP = 32
TOL = 1e-12
E8M0_BIAS = 127
Expand Down
5 changes: 5 additions & 0 deletions op_tests/triton_tests/rope/test_rope_mxfp4.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
)
from aiter.ops.triton.fusions.k_norm_rope_mxfp4_cache import k_norm_rope_mxfp4_cache
from aiter.ops.triton.rope.q_rope_mxfp4_quant import q_rope_mxfp4_quant
from aiter.ops.triton.utils._triton import arch_info
from op_tests.triton_tests.utils.pa_mqa_logits_mxfp4_ref import (
unshuffle_scales,
unshuffle_values,
Expand Down Expand Up @@ -244,6 +245,10 @@ def test_q_quant_matches_reference(num_tokens, num_heads, with_weights):


@pytest.mark.parametrize("page_size", [64, 128])
@pytest.mark.skipif(
arch_info.get_arch() != "gfx950",
reason="paged MXFP4 MQA logits is gfx950 only",
)
def test_round_trip_through_logits(page_size):
"""Keys written by the cache op and a query from the quant op score as the
dequantized values do: the writer and paged_mxfp4_mqa_logits agree."""
Expand Down
Loading