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

[CI] Reorganize stage-b 1-GPU tests for 5090 compatibility (#16826)

Browse Source
This commit is contained in:
Alison Shao
2026-01-15 15:23:35 -08:00
committed by GitHub
parent 8b22deef5b
commit 146b5fcc84
136 changed files with 277 additions and 404 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/lora/test_lora_backend.py
-1
View File
@@ -30,7 +30,6 @@ from sglang.test.lora_utils import (
from sglang.test.test_utils import CustomTestCase, is_in_ci
register_cuda_ci(est_time=200, suite="stage-b-test-small-1-gpu")
register_cuda_ci(est_time=137, suite="stage-b-test-small-1-gpu-5090")
register_amd_ci(
est_time=200,
suite="stage-b-test-small-1-gpu-amd",
test/registered/lora/test_lora_eviction.py
-1
View File
@@ -24,7 +24,6 @@ from sglang.test.runners import SRTRunner
from sglang.test.test_utils import CustomTestCase
register_cuda_ci(est_time=224, suite="stage-b-test-small-1-gpu")
register_cuda_ci(est_time=215, suite="stage-b-test-small-1-gpu-5090")
register_amd_ci(est_time=224, suite="stage-b-test-small-1-gpu-amd")
PROMPTS = [
test/registered/lora/test_lora_hf_sgl_logprob_diff.py
+1 -7
View File
@@ -36,6 +36,7 @@ import torch
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
from sglang.test.runners import HFRunner, SRTRunner
from sglang.test.test_utils import DEFAULT_PORT_FOR_SRT_TEST_RUNNER, CustomTestCase
register_cuda_ci(
est_time=150,
@@ -45,13 +46,6 @@ register_amd_ci(
est_time=250,
suite="stage-b-test-small-1-gpu-amd",
)
register_cuda_ci(
est_time=76,
suite="stage-b-test-small-1-gpu-5090",
)
from sglang.test.test_utils import DEFAULT_PORT_FOR_SRT_TEST_RUNNER, CustomTestCase
# Test configuration constants
LORA_BACKEND = "triton"
DISABLE_CUDA_GRAPH = False
test/registered/lora/test_lora_openai_compatible.py
+4 -5
View File
@@ -17,12 +17,8 @@ import unittest
import openai
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=150, suite="nightly-1-gpu", nightly=True)
register_amd_ci(est_time=150, suite="nightly-amd-1-gpu", nightly=True)
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
from sglang.test.test_utils import (
DEFAULT_SMALL_MODEL_NAME_FOR_TEST,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
@@ -31,6 +27,9 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=150, suite="nightly-1-gpu", nightly=True)
register_amd_ci(est_time=150, suite="nightly-amd-1-gpu", nightly=True)
def get_real_lora_adapter() -> str:
"""Use a real LoRA adapter from Hugging Face."""
test/registered/lora/test_lora_qwen3.py
+2 -3
View File
@@ -20,15 +20,14 @@ from sglang.test.lora_utils import (
LORA_MODELS_QWEN3,
run_lora_multiple_batch_on_model_cases,
)
from sglang.test.test_utils import CustomTestCase
register_cuda_ci(est_time=97, suite="nightly-1-gpu", nightly=True)
register_amd_ci(
est_time=30,
suite="stage-b-test-small-1-gpu-amd",
disabled="see https://github.com/sgl-project/sglang/issues/13107",
)
from sglang.test.test_utils import CustomTestCase
register_cuda_ci(est_time=97, suite="nightly-1-gpu", nightly=True)
class TestLoRAQwen3(CustomTestCase):
test/registered/lora/test_lora_update.py
+1 -1
View File
@@ -34,7 +34,7 @@ from sglang.test.test_utils import (
popen_launch_server,
)
register_cuda_ci(est_time=487, suite="stage-b-test-small-1-gpu")
register_cuda_ci(est_time=487, suite="stage-b-test-large-1-gpu")
PROMPTS = [
"SGL is a",
test/registered/lora/test_multi_lora_backend.py
+1 -1
View File
@@ -25,7 +25,7 @@ from sglang.test.lora_utils import (
)
from sglang.test.test_utils import CustomTestCase, is_in_ci
register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu")
register_cuda_ci(est_time=100, suite="stage-b-test-large-1-gpu")
register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu-amd")