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

ci: decouple stage and runner for cuda registry (#25197)

Browse Source
This commit is contained in:
Liangsheng Yin
2026-05-13 17:28:21 -07:00
committed by GitHub
parent dd98153d17
commit 22d3f3996c
261 changed files with 388 additions and 293 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/spec/dflash/test_dflash.py
+1 -1
View File
@@ -18,7 +18,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=302, suite="stage-b-test-1-gpu-small")
register_cuda_ci(est_time=302, stage="stage-b", runner_config="1-gpu-small")
class TestDFlashServerBase(CustomTestCase, MatchedStopMixin, GSM8KMixin):
test/registered/spec/eagle/test_adaptive_speculative.py
+1 -1
View File
@@ -18,7 +18,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=76, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=76, stage="stage-b", runner_config="1-gpu-large")
HIGH_ACCEPT_PROMPT = (
"Output exactly 128 new lines. "
test/registered/spec/eagle/test_deepseek_v3_fp4_mtp_small.py
+1 -1
View File
@@ -16,7 +16,7 @@ from sglang.test.test_utils import (
write_github_step_summary,
)
register_cuda_ci(est_time=420, suite="stage-b-test-4-gpu-b200")
register_cuda_ci(est_time=420, stage="stage-b", runner_config="4-gpu-b200")
FULL_DEEPSEEK_V3_FP4_MODEL_PATH = "nvidia/DeepSeek-V3-0324-FP4"
SERVER_LAUNCH_TIMEOUT = 1200
test/registered/spec/eagle/test_eagle_constrained_decoding.py
+1 -1
View File
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=116, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=116, stage="stage-b", runner_config="1-gpu-large")
class TestEagleConstrainedDecoding(
test/registered/spec/eagle/test_eagle_dp_attention.py
+1 -1
View File
@@ -21,7 +21,7 @@ from sglang.test.test_utils import (
)
# EAGLE3 with DP attention (tp=2, dp=2, requires 4 GPUs)
register_cuda_ci(est_time=99, suite="stage-c-test-4-gpu-h100")
register_cuda_ci(est_time=99, stage="stage-c", runner_config="4-gpu-h100")
register_amd_ci(est_time=200, suite="stage-c-test-4-gpu-amd")
test/registered/spec/eagle/test_eagle_infer_a.py
+1 -1
View File
@@ -13,7 +13,7 @@ from sglang.test.test_utils import (
CustomTestCase,
)
register_cuda_ci(est_time=357, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=357, stage="stage-b", runner_config="1-gpu-large")
class TestEAGLEEngine(CustomTestCase):
test/registered/spec/eagle/test_eagle_infer_b.py
+1 -1
View File
@@ -22,7 +22,7 @@ from sglang.test.run_eval import run_eval
from sglang.test.server_fixtures.eagle_fixture import EagleServerBase
from sglang.test.test_utils import DEFAULT_TARGET_MODEL_EAGLE, run_logprob_check
register_cuda_ci(est_time=847, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=847, stage="stage-b", runner_config="1-gpu-large")
class TestEAGLEServerBasic(EagleServerBase):
test/registered/spec/eagle/test_eagle_infer_beta.py
+1 -1
View File
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=369, suite="stage-b-test-1-gpu-small")
register_cuda_ci(est_time=369, stage="stage-b", runner_config="1-gpu-small")
class TestEagle3ServerBase(CustomTestCase, MatchedStopMixin):
test/registered/spec/eagle/test_eagle_infer_beta_dp_attention.py
+1 -1
View File
@@ -17,7 +17,7 @@ from sglang.test.test_utils import (
)
# EAGLE with DP attention on B200 (tp=2, dp=2, requires 4 B200 GPUs)
register_cuda_ci(est_time=123, suite="stage-c-test-4-gpu-b200")
register_cuda_ci(est_time=123, stage="stage-c", runner_config="4-gpu-b200")
def test_gsm8k(base_url: str, model: str):
test/registered/spec/test_constrained_decoding_spec_reasoning.py
+1 -1
View File
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
)
# Constrained decoding with EAGLE3 speculative reasoning (tp=2)
register_cuda_ci(est_time=137, suite="stage-b-test-2-gpu-large")
register_cuda_ci(est_time=137, stage="stage-b", runner_config="2-gpu-large")
class ServerWithGrammar(CustomTestCase):
test/registered/spec/test_ngram_speculative_decoding.py
+1 -1
View File
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=254, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=254, stage="stage-b", runner_config="1-gpu-large")
GSM_DATASET_PATH = None
test/registered/spec/test_standalone_speculative_decoding.py
+1 -1
View File
@@ -18,7 +18,7 @@ from sglang.test.test_utils import (
)
# Standalone speculative decoding tests (FA3, Triton, FlashInfer backends)
register_cuda_ci(est_time=406, suite="stage-b-test-1-gpu-large")
register_cuda_ci(est_time=406, stage="stage-b", runner_config="1-gpu-large")
GSM_DATASET_PATH = None
test/registered/spec/utils/test_build_eagle_tree.py
+1 -1
View File
@@ -9,7 +9,7 @@ from sglang.srt.speculative.eagle_utils import (
from sglang.srt.utils import get_device
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=6, suite="stage-b-test-1-gpu-small")
register_cuda_ci(est_time=6, stage="stage-b", runner_config="1-gpu-small")
register_amd_ci(est_time=3, suite="stage-b-test-1-gpu-small-amd")