Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 0 additions & 9 deletions jenkins/L0_Test.groovy
Original file line number Diff line number Diff line change
Expand Up @@ -5073,15 +5073,6 @@ def launchTestJobs(pipeline, testFilter)
36,
9
)
// 10 Nodes
multiNodesSBSAConfigs += buildStageConfigs(
"GB200-40_GPUs-10_Nodes-PyTorch-Disagg-PerfSanity-CTX1-NODE2-GPU8-GEN1-NODE8-GPU32-Post-Merge",
"auto:gb200-flex",
"l0_gb200_multi_nodes_perf_sanity_ctx1_node2_gpu8_gen1_node8_gpu32",
1,
40,
10
)
// GB300 PerfSanity post-merge aggregated
// 2 Nodes
multiNodesSBSAConfigs += buildStageConfigs(
Expand Down
2 changes: 0 additions & 2 deletions tests/integration/test_lists/qa/llm_perf_disagg.yml
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,6 @@ llm_perf_disagg:
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-UCX] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con1_ctx1_pp8_gen1_tep8_eplb0_mtp3_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con64_ctx1_pp8_gen1_dep32_eplb0_mtp3_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_ctx1_pp4_gen8_pp4_bs2_eplb0_mtp0_con2-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_1k1k_con1024_ctx1_dep4_gen1_dep32_eplb0_mtp0_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_1k1k_con1024_ctx1_dep4_gen1_dep32_eplb0_mtp3_ccb-NIXL] TIMEOUT (120)
Expand All @@ -37,7 +36,6 @@ llm_perf_disagg:
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con2048_ctx1_dep4_gen1_dep32_eplb288_mtp1_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con2048_ctx1_dep4_gen1_dep32_eplb288_mtp1_ccb-UCX] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con256_ctx1_dep4_gen1_dep32_eplb0_mtp3_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con256_ctx1_dep8_gen1_dep8_eplb0_mtp0_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con1024_ctx1_dep4_gen1_dep32_eplb256_mtp3_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con1_ctx1_dep4_gen1_tep8_eplb0_mtp3_ccb-NIXL] TIMEOUT (120)
- perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con4096_ctx1_dep4_gen1_dep32_eplb256_mtp0_ccb-NIXL] TIMEOUT (120)
Expand Down
2 changes: 0 additions & 2 deletions tests/integration/test_lists/qa/llm_perf_multinode.txt
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,6 @@
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-UCX]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con1_ctx1_pp8_gen1_tep8_eplb0_mtp3_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_con64_ctx1_pp8_gen1_dep32_eplb0_mtp3_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_128k8k_ctx1_pp4_gen8_pp4_bs2_eplb0_mtp0_con2-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_1k1k_con1024_ctx1_dep4_gen1_dep32_eplb0_mtp0_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-r1-fp4_1k1k_con1024_ctx1_dep4_gen1_dep32_eplb0_mtp3_ccb-NIXL]
Expand All @@ -31,7 +30,6 @@ perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con1_
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con2048_ctx1_dep4_gen1_dep32_eplb288_mtp1_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con2048_ctx1_dep4_gen1_dep32_eplb288_mtp1_ccb-UCX]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con256_ctx1_dep4_gen1_dep32_eplb0_mtp3_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_32k4k_con256_ctx1_dep8_gen1_dep8_eplb0_mtp0_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con1024_ctx1_dep4_gen1_dep32_eplb256_mtp3_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con1_ctx1_dep4_gen1_tep8_eplb0_mtp3_ccb-NIXL]
perf/test_perf_sanity.py::test_e2e[disagg-e2e-gb200_deepseek-v32-fp4_8k1k_con4096_ctx1_dep4_gen1_dep32_eplb256_mtp0_ccb-NIXL]
Expand Down

This file was deleted.

Original file line number Diff line number Diff line change
Expand Up @@ -29,4 +29,3 @@ l0_gb200_multi_nodes_perf_sanity_node2_gpu8:
# deepseek-r1-fp4
- perf/test_perf_sanity.py::test_e2e[aggr_upload-ctx_only-gb200_deepseek-r1-fp4_128k8k_con128_ctx1_pp8_gen1_dep16_eplb0_mtp1_ccb-NIXL] TIMEOUT (180)
- perf/test_perf_sanity.py::test_e2e[aggr_upload-ctx_only-gb200_deepseek-r1-fp4_128k8k_con1_ctx1_pp8_gen1_tep8_eplb0_mtp3_ccb-NIXL] TIMEOUT (180)
# - perf/test_perf_sanity.py::test_e2e[aggr_upload-ctx_only-gb200_deepseek-r1-fp4_128k8k_con64_ctx1_pp8_gen1_dep32_eplb0_mtp3_ccb-NIXL] TIMEOUT (180)

This file was deleted.

Original file line number Diff line number Diff line change
Expand Up @@ -52,13 +52,13 @@ worker_config:
pipeline_parallel_size: 1
context_parallel_size: 1
enable_attention_dp: true
max_batch_size: 8
max_batch_size: 32
max_num_tokens: 8192
max_seq_len: 121000
allreduce_strategy: MNNVL
cuda_graph_config:
enable_padding: true
max_batch_size: 8
max_batch_size: 32
kv_cache_config:
enable_block_reuse: false
free_gpu_memory_fraction: 0.9
Expand Down
Loading