minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[AMD] Register 5 JIT kernel unit tests for AMD nightly CI (#29197)

Browse Source
This commit is contained in:
Michael
2026-06-25 13:12:58 -07:00
committed by GitHub
parent 118d6b2e5e
commit e6efe10072
5 changed files with 10 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/jit/deepseek_v4/test_c128_v2.py
+2 -1
View File
@@ -16,10 +16,11 @@ from sglang.jit_kernel.tests.deepseek_v4.common import (
make_state_pool, make_state_pool,
to_seq_extend, to_seq_extend,
) )
from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=30, suite="base-b-kernel-unit-1-gpu-large") register_cuda_ci(est_time=30, suite="base-b-kernel-unit-1-gpu-large")
register_cuda_ci(est_time=30, suite="nightly-kernel-1-gpu", nightly=True) register_cuda_ci(est_time=30, suite="nightly-kernel-1-gpu", nightly=True)
register_amd_ci(est_time=30, suite="nightly-amd-kernel-1-gpu", nightly=True)
Context = Union[LegacyContext, PagedContext] Context = Union[LegacyContext, PagedContext]
test/registered/jit/deepseek_v4/test_c4_v2.py
+2 -1
View File
@@ -16,10 +16,11 @@ from sglang.jit_kernel.tests.deepseek_v4.common import (
make_state_pool, make_state_pool,
to_seq_extend, to_seq_extend,
) )
from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=30, suite="base-b-kernel-unit-1-gpu-large") register_cuda_ci(est_time=30, suite="base-b-kernel-unit-1-gpu-large")
register_cuda_ci(est_time=30, suite="nightly-kernel-1-gpu", nightly=True) register_cuda_ci(est_time=30, suite="nightly-kernel-1-gpu", nightly=True)
register_amd_ci(est_time=30, suite="nightly-amd-kernel-1-gpu", nightly=True)
Context = Union[LegacyContext, PagedContext] Context = Union[LegacyContext, PagedContext]
test/registered/jit/diffusion/test_group_norm_silu.py
+2 -1
View File
@@ -7,10 +7,11 @@ import torch.nn.functional as F
from sglang.jit_kernel.diffusion.group_norm_silu import apply_group_norm_silu from sglang.jit_kernel.diffusion.group_norm_silu import apply_group_norm_silu
from sglang.jit_kernel.diffusion.triton.group_norm_silu import triton_group_norm_silu from sglang.jit_kernel.diffusion.triton.group_norm_silu import triton_group_norm_silu
from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=8, suite="base-b-kernel-unit-1-gpu-large") register_cuda_ci(est_time=8, suite="base-b-kernel-unit-1-gpu-large")
register_cuda_ci(est_time=120, suite="nightly-kernel-1-gpu", nightly=True) register_cuda_ci(est_time=120, suite="nightly-kernel-1-gpu", nightly=True)
register_amd_ci(est_time=15, suite="nightly-amd-kernel-1-gpu", nightly=True)
DEVICE = "cuda" DEVICE = "cuda"
DTYPES = [torch.float16, torch.bfloat16, torch.float32] DTYPES = [torch.float16, torch.bfloat16, torch.float32]
test/registered/jit/diffusion/test_qwen_image_modulation.py
+2 -1
View File
@@ -10,10 +10,11 @@ from sglang.jit_kernel.diffusion.triton.scale_shift import (
fuse_residual_layernorm_scale_shift_gate_select01_kernel, fuse_residual_layernorm_scale_shift_gate_select01_kernel,
) )
from sglang.jit_kernel.utils import get_ci_test_range from sglang.jit_kernel.utils import get_ci_test_range
from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=15, suite="base-b-kernel-unit-1-gpu-large") register_cuda_ci(est_time=15, suite="base-b-kernel-unit-1-gpu-large")
register_cuda_ci(est_time=120, suite="nightly-kernel-1-gpu", nightly=True) register_cuda_ci(est_time=120, suite="nightly-kernel-1-gpu", nightly=True)
register_amd_ci(est_time=30, suite="nightly-amd-kernel-1-gpu", nightly=True)
DEVICE = "cuda" DEVICE = "cuda"
DTYPES = get_ci_test_range( DTYPES = get_ci_test_range(
test/registered/jit/diffusion/test_varlen_pack_pad.py
+2 -1
View File
@@ -13,10 +13,11 @@ from sglang.jit_kernel.diffusion.triton.varlen_pack_pad import (
fused_scatter_to_padded, fused_scatter_to_padded,
) )
from sglang.jit_kernel.utils import get_ci_test_range from sglang.jit_kernel.utils import get_ci_test_range
from sglang.test.ci.ci_register import register_cuda_ci from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=10, suite="base-b-kernel-unit-1-gpu-large") register_cuda_ci(est_time=10, suite="base-b-kernel-unit-1-gpu-large")
register_cuda_ci(est_time=60, suite="nightly-kernel-1-gpu", nightly=True) register_cuda_ci(est_time=60, suite="nightly-kernel-1-gpu", nightly=True)
register_amd_ci(est_time=15, suite="nightly-amd-kernel-1-gpu", nightly=True)
DEVICE = "cuda" DEVICE = "cuda"
DTYPES = get_ci_test_range([torch.bfloat16, torch.float16], [torch.bfloat16]) DTYPES = get_ci_test_range([torch.bfloat16, torch.float16], [torch.bfloat16])