kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-02-04 18:21:52 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

[TRTINFRA-7548][infra] Update GB200 test configs to use frontend SLURM platforms (#11085)

Browse Source 
Signed-off-by: Matt Lefebvre <mlefebvre@nvidia.com>
This commit is contained in:
Matt Lefebvre 2026-01-30 14:07:47 -08:00 committed by GitHub
parent 5a97374f3c
commit 97ab014bdb
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 21 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
jenkins/L0_Test.groovy
View File

@ -3330,40 +3330,40 @@ def launchTestJobs(pipeline, testFilter)
fullSet += SBSATestConfigs.keySet()
SBSASlurmTestConfigs = [
"GB200-4_GPUs-PyTorch-1": ["gb200-x4-oci", "l0_gb200_multi_gpus", 1, 2, 4],
"GB200-4_GPUs-PyTorch-2": ["gb200-x4-oci", "l0_gb200_multi_gpus", 2, 2, 4],
"GB200-4_GPUs-PyTorch-Post-Merge-1": ["gb200-x4-oci", "l0_gb200_multi_gpus", 1, 1, 4],
"GB200-4_GPUs-PyTorch-1": ["auto:gb200-x4", "l0_gb200_multi_gpus", 1, 2, 4],
"GB200-4_GPUs-PyTorch-2": ["auto:gb200-x4", "l0_gb200_multi_gpus", 2, 2, 4],
"GB200-4_GPUs-PyTorch-Post-Merge-1": ["auto:gb200-x4", "l0_gb200_multi_gpus", 1, 1, 4],
"GB10-PyTorch-Post-Merge-1": ["gb10x-single", "l0_gb10", 1, 1],
// Disable GB300 stages due to nodes will be offline temporarily.
// "GB300-PyTorch-1": ["gb300-single", "l0_gb300", 1, 1],
// "GB300-4_GPUs-PyTorch-Post-Merge-1": ["gb300-x4", "l0_gb300_multi_gpus", 1, 1, 4],
// PerfSanity pre-merge tests
"GB200-4_GPUs-PyTorch-PerfSanity-1": ["gb200-x4-oci", "l0_gb200_multi_gpus_perf_sanity", 1, 1, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-1": ["auto:gb200-x4", "l0_gb200_multi_gpus_perf_sanity", 1, 1, 4],
// PerfSanity post-merge tests
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-1": ["gb200-x4-oci", "l0_gb200_multi_gpus_perf_sanity", 1, 3, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-2": ["gb200-x4-oci", "l0_gb200_multi_gpus_perf_sanity", 2, 3, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-3": ["gb200-x4-oci", "l0_gb200_multi_gpus_perf_sanity", 3, 3, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-1": ["auto:gb200-x4", "l0_gb200_multi_gpus_perf_sanity", 1, 3, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-2": ["auto:gb200-x4", "l0_gb200_multi_gpus_perf_sanity", 2, 3, 4],
"GB200-4_GPUs-PyTorch-PerfSanity-Post-Merge-3": ["auto:gb200-x4", "l0_gb200_multi_gpus_perf_sanity", 3, 3, 4],
]
fullSet += SBSASlurmTestConfigs.keySet()
multiNodesSBSAConfigs = [
// Each testcase uses 8 GPUs and 2 nodes.
// https://nvbugs/5598863 (uncorrectable NVLink error detected during the execution) may not exist in OCI machines.
"GB200-8_GPUs-2_Nodes-PyTorch-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes", 1, 2, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-2": ["gb200-oci-trtllm", "l0_gb200_multi_nodes", 2, 2, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes", 1, 3, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-2": ["gb200-oci-trtllm", "l0_gb200_multi_nodes", 2, 3, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-3": ["gb200-oci-trtllm", "l0_gb200_multi_nodes", 3, 3, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-1": ["auto:gb200-flex", "l0_gb200_multi_nodes", 1, 2, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-2": ["auto:gb200-flex", "l0_gb200_multi_nodes", 2, 2, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-1": ["auto:gb200-flex", "l0_gb200_multi_nodes", 1, 3, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-2": ["auto:gb200-flex", "l0_gb200_multi_nodes", 2, 3, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-Post-Merge-3": ["auto:gb200-flex", "l0_gb200_multi_nodes", 3, 3, 8, 2],
// PerfSanity post-merge tests
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 1, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-2": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 2, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-3": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 3, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-4": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 4, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-5": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 5, 5, 8, 2],
"GB200-12_GPUs-3_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_disagg_perf_sanity_3_nodes", 1, 1, 12, 3],
// "GB200-24_GPUs-6_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_disagg_perf_sanity_6_nodes", 1, 2, 24, 6],
// "GB200-24_GPUs-6_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-2": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_disagg_perf_sanity_6_nodes", 2, 2, 24, 6],
// "GB200-32_GPUs-8_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["gb200-oci-trtllm", "l0_gb200_multi_nodes_disagg_perf_sanity_8_nodes", 1, 1, 32, 8],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-1": ["auto:gb200-flex", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 1, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-2": ["auto:gb200-flex", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 2, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-3": ["auto:gb200-flex", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 3, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-4": ["auto:gb200-flex", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 4, 5, 8, 2],
"GB200-8_GPUs-2_Nodes-PyTorch-PerfSanity-Post-Merge-5": ["auto:gb200-flex", "l0_gb200_multi_nodes_aggr_perf_sanity_2_nodes", 5, 5, 8, 2],
"GB200-12_GPUs-3_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["auto:gb200-flex", "l0_gb200_multi_nodes_disagg_perf_sanity_3_nodes", 1, 1, 12, 3],
// "GB200-24_GPUs-6_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["auto:gb200-flex", "l0_gb200_multi_nodes_disagg_perf_sanity_6_nodes", 1, 2, 24, 6],
// "GB200-24_GPUs-6_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-2": ["auto:gb200-flex", "l0_gb200_multi_nodes_disagg_perf_sanity_6_nodes", 2, 2, 24, 6],
// "GB200-32_GPUs-8_Nodes-PyTorch-PerfSanity-Disagg-Post-Merge-1": ["auto:gb200-flex", "l0_gb200_multi_nodes_disagg_perf_sanity_8_nodes", 1, 1, 32, 8],
]
fullSet += multiNodesSBSAConfigs.keySet()