[ROCm][CI] Set all timeout_in_minutes to 180 (#48146)
Signed-off-by: charlifu <charlifu@amd.com>
Charlie Fu committed
Jul 9, 2026 at 12:52 UTC
67e7ea8977bc4281d4d33dd3a81a5c5fab3df920
1 file changed
+27
-27
.buildkite/test-amd.yaml
+27
-27
@@ -275,7 +275,7 @@ steps:
275
- pytest -v -s v1/cudagraph/test_cudagraph_mode.py
276
277
- label: e2e Core (1 GPU) # TBD
278
- timeout_in_minutes: 35
278
+ timeout_in_minutes: 180
279
mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250]
280
agent_pool: mi250_1
281
optional: true
@@ -439,7 +439,7 @@ steps:
439
#----------------------------------------------------------- mi250 · docker ----------------------------------------------------------#
440
441
- label: Docker Build Metadata (ROCm) # TBD
442
- timeout_in_minutes: 20
442
+ timeout_in_minutes: 180
443
mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250]
444
agent_pool: mi250_1
445
no_gpu: true
@@ -467,7 +467,7 @@ steps:
467
#----------------------------------------------------- mi300 · basic_correctness -----------------------------------------------------#
468
469
- label: Basic Correctness # TBD
470
- timeout_in_minutes: 95
470
+ timeout_in_minutes: 180
471
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
472
agent_pool: mi300_1
473
fast_check: true
@@ -484,7 +484,7 @@ steps:
484
- pytest -v -s basic_correctness/test_cpu_offload.py
485
486
- label: Distributed Model Tests (2 GPUs) # TBD
487
- timeout_in_minutes: 110
487
+ timeout_in_minutes: 180
488
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
489
agent_pool: mi300_2
490
num_gpus: 2
@@ -724,7 +724,7 @@ steps:
724
- pytest -v -s distributed/test_eplb_spec_decode.py
725
726
- label: Distributed Tests (2xH100-2xMI300) # TBD
727
- timeout_in_minutes: 75
727
+ timeout_in_minutes: 180
728
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
729
agent_pool: mi300_2
730
num_gpus: 2
@@ -1297,7 +1297,7 @@ steps:
1297
#--------------------------------------------------------- mi300 · examples ----------------------------------------------------------#
1298
1299
- label: Examples # TBD
1300
- timeout_in_minutes: 90
1300
+ timeout_in_minutes: 180
1301
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1302
agent_pool: mi300_1
1303
optional: true
@@ -1348,7 +1348,7 @@ steps:
1348
- pytest -v -s tests/kernels/ir
1349
1350
- label: Kernels Attention Test %N # TBD
1351
- timeout_in_minutes: 100
1351
+ timeout_in_minutes: 180
1352
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1353
agent_pool: mi300_1
1354
optional: true
@@ -1397,7 +1397,7 @@ steps:
1397
- pytest -v -s kernels/test_kda.py
1398
1399
- label: Kernels MoE Test %N # TBD
1400
- timeout_in_minutes: 95
1400
+ timeout_in_minutes: 180
1401
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1402
agent_pool: mi300_1
1403
optional: true
@@ -1511,7 +1511,7 @@ steps:
1511
#---------------------------------------------------- mi300 · model_runner_v2 -------------------------------------------------------#
1512
1513
- label: Model Runner V2 Core Tests # TBD
1514
- timeout_in_minutes: 60
1514
+ timeout_in_minutes: 180
1515
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1516
agent_pool: mi300_1
1517
optional: true
@@ -1535,7 +1535,7 @@ steps:
1535
- pytest -v -s entrypoints/llm/test_struct_output_generate.py -k "xgrammar and not speculative_config6 and not speculative_config7 and not speculative_config8 and not speculative_config0"
1536
1537
- label: Model Runner V2 Examples # TBD
1538
- timeout_in_minutes: 60
1538
+ timeout_in_minutes: 180
1539
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1540
agent_pool: mi300_1
1541
optional: true
@@ -1569,7 +1569,7 @@ steps:
1569
- python3 features/speculative_decoding/spec_decode_offline.py --test --method eagle3 --num_spec_tokens 3 --dataset-name hf --dataset-path philschmid/mt-bench --num-prompts 80 --temp 0 --top-p 1.0 --top-k -1 --tp 1 --enable-chunked-prefill --max-model-len 1536
1570
1571
- label: Model Runner V2 Distributed (2 GPUs) # TBD
1572
- timeout_in_minutes: 60
1572
+ timeout_in_minutes: 180
1573
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1574
agent_pool: mi300_2
1575
num_gpus: 2
@@ -1661,7 +1661,7 @@ steps:
1661
- pytest -v -s models/test_initialization.py::test_can_initialize_small_subset
1662
1663
- label: Basic Models Tests (Other) # TBD
1664
- timeout_in_minutes: 90
1664
+ timeout_in_minutes: 180
1665
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1666
agent_pool: mi300_1
1667
optional: true
@@ -1806,7 +1806,7 @@ steps:
1806
- cd .. && VLLM_WORKER_MULTIPROC_METHOD=spawn pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model
1807
1808
- label: Multi-Modal Processor # 1h 42m
1809
- timeout_in_minutes: 138
1809
+ timeout_in_minutes: 180
1810
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1811
agent_pool: mi300_1
1812
optional: true
@@ -2255,7 +2255,7 @@ steps:
2255
- pytest -v -s v1/e2e/spec_decode -k "draft_model or no_sync or batch_inference"
2256
2257
- label: Spec Decode Eagle # TBD
2258
- timeout_in_minutes: 90
2258
+ timeout_in_minutes: 180
2259
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
2260
agent_pool: mi300_1
2261
optional: true
@@ -2472,7 +2472,7 @@ steps:
2472
- DP_SIZE=2 pytest -v -s entrypoints/openai/test_multi_api_servers.py
2473
2474
- label: Metrics, Tracing (2 GPUs) # TBD
2475
- timeout_in_minutes: 65
2475
+ timeout_in_minutes: 180
2476
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
2477
agent_pool: mi300_2
2478
optional: true
@@ -2678,7 +2678,7 @@ steps:
2678
#------------------------------------------------------ mi300 · weight_loading -------------------------------------------------------#
2679
2680
- label: Weight Loading Multiple GPU # TBD
2681
- timeout_in_minutes: 75
2681
+ timeout_in_minutes: 180
2682
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
2683
agent_pool: mi300_2
2684
num_gpus: 2
@@ -2690,7 +2690,7 @@ steps:
2690
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt
2691
2692
- label: Weight Loading Multiple GPU - Large Models # TBD
2693
- timeout_in_minutes: 75
2693
+ timeout_in_minutes: 180
2694
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
2695
agent_pool: mi300_2
2696
num_gpus: 2
@@ -3089,7 +3089,7 @@ steps:
3089
- pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-gfx950.txt
3090
3091
- label: LM Eval Qwen3-5 Models (B200-MI355) # TBD
3092
- timeout_in_minutes: 120
3092
+ timeout_in_minutes: 180
3093
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3094
agent_pool: mi355_2
3095
num_gpus: 2
@@ -3171,7 +3171,7 @@ steps:
3171
#--------------------------------------------------------- mi355 · examples ----------------------------------------------------------#
3172
3173
- label: Examples # TBD
3174
- timeout_in_minutes: 100
3174
+ timeout_in_minutes: 180
3175
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3176
agent_pool: mi355_1
3177
working_dir: "/vllm-workspace/examples"
@@ -3206,7 +3206,7 @@ steps:
3206
#---------------------------------------------------------- mi355 · kernels ----------------------------------------------------------#
3207
3208
- label: Kernels (B200-MI355) # TBD
3209
- timeout_in_minutes: 15
3209
+ timeout_in_minutes: 180
3210
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3211
agent_pool: mi355_1
3212
working_dir: "/vllm-workspace/"
@@ -3231,7 +3231,7 @@ steps:
3231
- pytest -v -s tests/kernels/attention/test_rocm_aiter_mla_decode_metadata.py
3232
3233
- label: Kernels Attention Test %N # TBD
3234
- timeout_in_minutes: 100
3234
+ timeout_in_minutes: 180
3235
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3236
agent_pool: mi355_1
3237
parallelism: 2
@@ -3248,7 +3248,7 @@ steps:
3248
- pytest -v -s kernels/attention --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
3249
3250
- label: Kernels MoE Test %N # TBD
3251
- timeout_in_minutes: 50
3251
+ timeout_in_minutes: 180
3252
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3253
agent_pool: mi355_1
3254
parallelism: 5
@@ -3512,7 +3512,7 @@ steps:
3512
- pytest -v -s v1/attention
3513
3514
- label: V1 Core + KV + Metrics # TBD
3515
- timeout_in_minutes: 60
3515
+ timeout_in_minutes: 180
3516
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3517
agent_pool: mi355_1
3518
optional: true
@@ -3538,7 +3538,7 @@ steps:
3538
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
3539
3540
- label: V1 Sample + Logits # TBD
3541
- timeout_in_minutes: 60
3541
+ timeout_in_minutes: 180
3542
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3543
agent_pool: mi355_1
3544
optional: true
@@ -3558,7 +3558,7 @@ steps:
3558
- pytest -v -s v1/test_outputs.py
3559
3560
- label: V1 Spec Decode # TBD
3561
- timeout_in_minutes: 60
3561
+ timeout_in_minutes: 180
3562
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3563
agent_pool: mi355_1
3564
working_dir: "/vllm-workspace/tests"
@@ -3571,7 +3571,7 @@ steps:
3571
#------------------------------------------------------ mi355 · weight_loading -------------------------------------------------------#
3572
3573
- label: Weight Loading Multiple GPU # TBD
3574
- timeout_in_minutes: 75
3574
+ timeout_in_minutes: 180
3575
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3576
agent_pool: mi355_2
3577
num_gpus: 2
@@ -3583,7 +3583,7 @@ steps:
3583
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt
3584
3585
- label: Weight Loading Multiple GPU - Large Models # TBD
3586
- timeout_in_minutes: 75
3586
+ timeout_in_minutes: 180
3587
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
3588
agent_pool: mi355_2
3589
working_dir: "/vllm-workspace/tests"