Skip to content

[CI] Register jit_kernel Test Files to Solve No Registry Found Error #21541

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

Closed
bingxche wants to merge 1 commit into from
Closed

[CI] Register jit_kernel Test Files to Solve No Registry Found Error #21541

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
7 changes: 7 additions & 0 deletions python/sglang/jit_kernel/benchmark/bench_cast.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,13 @@
run_benchmark,
)
from sglang.jit_kernel.cast import downcast_fp8 as downcast_fp8_jit
from sglang.test.ci.ci_register import register_cuda_ci

register_cuda_ci(
est_time=13,
suite="stage-b-kernel-benchmark-1-gpu-large",
disabled="unknown est_time",
)
Comment on lines +14 to +18
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot Mar 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The disabled parameter will cause this test to be registered but skipped by the CI. Since an est_time is provided and the goal of this PR is to enable these tests, the disabled parameter should be removed.

register_cuda_ci(
    est_time=13,
    suite="stage-b-kernel-benchmark-1-gpu-large",
)


DEVICE = DEFAULT_DEVICE
DTYPE = torch.bfloat16
Expand Down
7 changes: 7 additions & 0 deletions python/sglang/jit_kernel/benchmark/bench_fused_qknorm_rope.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,13 @@
from sglang.jit_kernel.fused_qknorm_rope import (
fused_qk_norm_rope as fused_qk_norm_rope_jit,
)
from sglang.test.ci.ci_register import register_cuda_ci

register_cuda_ci(
est_time=17,
suite="stage-b-kernel-benchmark-1-gpu-large",
disabled="unknown est_time",
)
Comment on lines +23 to +27
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot Mar 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The disabled parameter will cause this test to be registered but skipped by the CI. Since an est_time is provided and the goal of this PR is to enable these tests, the disabled parameter should be removed.

register_cuda_ci(
    est_time=17,
    suite="stage-b-kernel-benchmark-1-gpu-large",
)


try:
from sgl_kernel import fused_qk_norm_rope as fused_qk_norm_rope_aot
Expand Down
13 changes: 13 additions & 0 deletions python/sglang/jit_kernel/tests/test_cast.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,19 @@
import torch

from sglang.jit_kernel.cast import downcast_fp8
from sglang.test.ci.ci_register import register_cuda_ci

register_cuda_ci(
est_time=24,
suite="stage-b-kernel-unit-1-gpu-large",
disabled="unknown est_time",
)
register_cuda_ci(
est_time=120,
suite="nightly-kernel-1-gpu",
nightly=True,
disabled="unknown est_time",
)
Comment on lines +7 to +17
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot Mar 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The disabled parameter will cause these tests to be registered but skipped by the CI. Since an est_time is provided for both and the goal of this PR is to enable these tests, the disabled parameter should be removed from both register_cuda_ci calls.

register_cuda_ci(
    est_time=24,
    suite="stage-b-kernel-unit-1-gpu-large",
)
register_cuda_ci(
    est_time=120,
    suite="nightly-kernel-1-gpu",
    nightly=True,
)


DTYPES = [torch.bfloat16, torch.float16]

Expand Down
13 changes: 13 additions & 0 deletions python/sglang/jit_kernel/tests/test_fused_qknorm_rope.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,19 @@
import torch

from sglang.jit_kernel.fused_qknorm_rope import fused_qk_norm_rope
from sglang.test.ci.ci_register import register_cuda_ci

register_cuda_ci(
est_time=64,
suite="stage-b-kernel-unit-1-gpu-large",
disabled="unknown est_time",
)
register_cuda_ci(
est_time=256,
suite="nightly-kernel-1-gpu",
nightly=True,
disabled="unknown est_time",
)
Comment on lines +14 to +24
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot Mar 27, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

high

The disabled parameter will cause these tests to be registered but skipped by the CI. Since an est_time is provided for both and the goal of this PR is to enable these tests, the disabled parameter should be removed from both register_cuda_ci calls.

register_cuda_ci(
    est_time=64,
    suite="stage-b-kernel-unit-1-gpu-large",
)
register_cuda_ci(
    est_time=256,
    suite="nightly-kernel-1-gpu",
    nightly=True,
)


try:
from sgl_kernel import fused_qk_norm_rope as fused_qk_norm_rope_aot
Expand Down
Loading