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

fix: adding matrix partitioning for h200 and b200 nightly tests (#17091)

Browse Source
This commit is contained in:
Douglas Yang
2026-01-15 11:23:19 -08:00
committed by GitHub
parent 3f44268fe5
commit 655d2c7c2a
11 changed files with 21 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
test/registered/8-gpu-models/test_deepseek_v31.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=5400, suite="nightly-8-gpu-common", nightly=True)
DEEPSEEK_V31_MODEL_PATH = "deepseek-ai/DeepSeek-V3.1"

2
test/registered/8-gpu-models/test_deepseek_v32.py
View File

@@ -6,7 +6,7 @@ from sglang.test.performance_test_runner import PerformanceTestParams
from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings, is_blackwell_system
register_cuda_ci(est_time=18000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=5400, suite="nightly-8-gpu-common", nightly=True)
DEEPSEEK_V32_MODEL_PATH = "deepseek-ai/DeepSeek-V3.2"

2
test/registered/8-gpu-models/test_deepseek_v32_cp_single_node.py
View File

@@ -5,7 +5,7 @@ from sglang.test.ci.ci_register import register_cuda_ci
from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings, is_blackwell_system
register_cuda_ci(est_time=18000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=5400, suite="nightly-8-gpu-common", nightly=True)
DEEPSEEK_V32_EXP_MODEL_PATH = "deepseek-ai/DeepSeek-V3.2-Exp"

2
test/registered/8-gpu-models/test_glm_46.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
GLM_4_6_MODEL_PATH = "zai-org/GLM-4.6"

2
test/registered/8-gpu-models/test_glm_46_fp8.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
GLM_4_6_FP8_MODEL_PATH = "zai-org/GLM-4.6-FP8"

2
test/registered/8-gpu-models/test_kimi_k2.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
KIMI_K2_THINKING_MODEL_PATH = "moonshotai/Kimi-K2-Thinking"

2
test/registered/8-gpu-models/test_llama4.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
LLAMA4_MODEL_PATH = "meta-llama/Llama-4-Scout-17B-16E-Instruct"

2
test/registered/8-gpu-models/test_minimax_m2.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
MINIMAX_M2_MODEL_PATH = "MiniMaxAI/MiniMax-M2"

2
test/registered/8-gpu-models/test_mistral_large3.py
View File

@@ -9,7 +9,7 @@ from sglang.test.test_utils import ModelLaunchSettings, is_blackwell_system
# Runs on both H200 and B200 via nightly-8-gpu-common suite
# Note: trtllm_mla backend may have hardware-specific behavior
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
MISTRAL_LARGE3_MODEL_PATH = "mistralai/Mistral-Large-3-675B-Instruct-2512"
MISTRAL_LARGE3_EAGLE_MODEL_PATH = "mistralai/Mistral-Large-3-675B-Instruct-2512-Eagle"

2
test/registered/8-gpu-models/test_qwen3_235b.py
View File

@@ -7,7 +7,7 @@ from sglang.test.run_combined_tests import run_combined_tests
from sglang.test.test_utils import ModelLaunchSettings, is_blackwell_system
# Runs on both H200 and B200 via nightly-8-gpu-common suite
register_cuda_ci(est_time=12000, suite="nightly-8-gpu-common", nightly=True)
register_cuda_ci(est_time=1800, suite="nightly-8-gpu-common", nightly=True)
QWEN3_235B_MODEL_PATH = "Qwen/Qwen3-235B-A22B-Instruct-2507"