A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 23:45 UTC
Changing "amdproduction" allocation. (#21409)
Signed-off-by: Alexei V. Ivanov <alexei.ivanov@amd.com>
Alexei-V-Ivanov-AMD committed
Jul 22, 2025 at 22:48 UTC
107111a85991cbe9c8d16b697e048d91aa4135f6
1 file changed
+11
-11
.buildkite/test-pipeline.yaml
+11
-11
@@ -225,7 +225,7 @@ steps:
225
##### 1 GPU test #####
226
227
- label: Regression Test # 5min
228
- mirror_hardwares: [amdexperimental]
228
+ mirror_hardwares: [amdexperimental, amdproduction]
229
source_file_dependencies:
230
- vllm/
231
- tests/test_regression
@@ -277,7 +277,7 @@ steps:
277
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
278
279
- label: Examples Test # 25min
280
- mirror_hardwares: [amdexperimental]
280
+ mirror_hardwares: [amdexperimental, amdproduction]
281
working_dir: "/vllm-workspace/examples"
282
source_file_dependencies:
283
- vllm/entrypoints
@@ -311,7 +311,7 @@ steps:
311
312
313
- label: Platform Tests (CUDA)
314
- mirror_hardwares: [amdexperimental]
314
+ mirror_hardwares: [amdexperimental, amdproduction]
315
source_file_dependencies:
316
- vllm/
317
- tests/cuda
@@ -330,7 +330,7 @@ steps:
330
- VLLM_USE_FLASHINFER_SAMPLER=1 pytest -v -s samplers
331
332
- label: LoRA Test %N # 15min each
333
- mirror_hardwares: [amdexperimental, amdproduction]
333
+ mirror_hardwares: [amdexperimental]
334
source_file_dependencies:
335
- vllm/lora
336
- tests/lora
@@ -382,7 +382,7 @@ steps:
382
- pytest -v -s kernels/core
383
384
- label: Kernels Attention Test %N
385
- mirror_hardwares: [amdexperimental, amdproduction]
385
+ mirror_hardwares: [amdexperimental]
386
source_file_dependencies:
387
- csrc/attention/
388
- vllm/attention
@@ -393,7 +393,7 @@ steps:
393
parallelism: 2
394
395
- label: Kernels Quantization Test %N
396
- mirror_hardwares: [amdexperimental, amdproduction]
396
+ mirror_hardwares: [amdexperimental]
397
source_file_dependencies:
398
- csrc/quantization/
399
- vllm/model_executor/layers/quantization
@@ -412,7 +412,7 @@ steps:
412
- pytest -v -s kernels/moe
413
414
- label: Kernels Mamba Test
415
- mirror_hardwares: [amdexperimental]
415
+ mirror_hardwares: [amdexperimental, amdproduction]
416
source_file_dependencies:
417
- csrc/mamba/
418
- tests/kernels/mamba
@@ -420,7 +420,7 @@ steps:
420
- pytest -v -s kernels/mamba
421
422
- label: Tensorizer Test # 11min
423
- mirror_hardwares: [amdexperimental]
423
+ mirror_hardwares: [amdexperimental, amdproduction]
424
soft_fail: true
425
source_file_dependencies:
426
- vllm/model_executor/model_loader
@@ -490,7 +490,7 @@ steps:
490
- pytest -s entrypoints/openai/correctness/
491
492
- label: Encoder Decoder tests # 5min
493
- mirror_hardwares: [amdexperimental]
493
+ mirror_hardwares: [amdexperimental, amdproduction]
494
source_file_dependencies:
495
- vllm/
496
- tests/encoder_decoder
@@ -498,7 +498,7 @@ steps:
498
- pytest -v -s encoder_decoder
499
500
- label: OpenAI-Compatible Tool Use # 20 min
501
- mirror_hardwares: [amdexperimental]
501
+ mirror_hardwares: [amdexperimental, amdproduction]
502
fast_check: false
503
source_file_dependencies:
504
- vllm/
@@ -610,7 +610,7 @@ steps:
610
- pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model'
611
612
- label: Quantized Models Test
613
- mirror_hardwares: [amdexperimental, amdproduction]
613
+ mirror_hardwares: [amdexperimental]
614
source_file_dependencies:
615
- vllm/model_executor/layers/quantization
616
- tests/models/quantization