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 2, 14:45 UTC
[ROCm][CI] Move workload from MI300 to MI325 (#43824)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Andreas Karatzas committed
May 28, 2026 at 05:31 UTC
a9bc0ad8e49c182aaa1a07c6610c884ecabc56a1
6 files changed
+14
-14
.buildkite/test_areas/engine.yaml
+1
-1
@@ -38,7 +38,7 @@ steps:
38
- pytest -v -s v1/engine --ignore v1/engine/test_preprocess_error_handling.py
39
mirror:
40
amd:
41
- device: mi300_1
41
+ device: mi325_1
42
timeout_in_minutes: 40
43
depends_on:
44
- image-build-amd
.buildkite/test_areas/entrypoints.yaml
+5
-5
@@ -28,7 +28,7 @@ steps:
28
- pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
29
mirror:
30
amd:
31
- device: mi300_1
31
+ device: mi325_1
32
soft_fail: true
33
depends_on:
34
- image-build-amd
@@ -46,7 +46,7 @@ steps:
46
- pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_chat_with_tool_reasoning.py --ignore=entrypoints/openai/chat_completion/test_oot_registration.py
47
mirror:
48
amd:
49
- device: mi300_1
49
+ device: mi325_1
50
soft_fail: true
51
timeout_in_minutes: 80
52
depends_on:
@@ -65,7 +65,7 @@ steps:
65
- pytest -v -s entrypoints/test_chat_utils.py
66
mirror:
67
amd:
68
- device: mi300_1
68
+ device: mi325_1
69
soft_fail: true
70
timeout_in_minutes: 60
71
depends_on:
@@ -85,7 +85,7 @@ steps:
85
- pytest -v -s entrypoints/openai --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/tool_parsers/ --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/test_multi_api_servers.py
86
mirror:
87
amd:
88
- device: mi300_1
88
+ device: mi325_1
89
soft_fail: true
90
timeout_in_minutes: 60
91
depends_on:
@@ -108,7 +108,7 @@ steps:
108
- pytest -v -s tool_use
109
mirror:
110
amd:
111
- device: mi300_1
111
+ device: mi325_1
112
soft_fail: true
113
depends_on:
114
- image-build-amd
.buildkite/test_areas/kernels.yaml
+1
-1
@@ -86,7 +86,7 @@ steps:
86
parallelism: 2
87
mirror:
88
amd:
89
- device: mi300_1
89
+ device: mi325_1
90
source_file_dependencies:
91
- csrc/quantization/
92
- vllm/model_executor/layers/quantization
.buildkite/test_areas/misc.yaml
+1
-1
@@ -52,7 +52,7 @@ steps:
52
- pytest -v -s v1/test_outputs.py
53
mirror:
54
amd:
55
- device: mi300_1
55
+ device: mi325_1
56
depends_on:
57
- image-build-amd
58
.buildkite/test_areas/models_language.yaml
+2
-2
@@ -50,7 +50,7 @@ steps:
50
mirror:
51
torch_nightly: {}
52
amd:
53
- device: mi300_1
53
+ device: mi325_1
54
depends_on:
55
- image-build-amd
56
commands:
@@ -96,7 +96,7 @@ steps:
96
- pytest -v -s models/language/pooling -m 'not core_model'
97
mirror:
98
amd:
99
- device: mi300_1
99
+ device: mi325_1
100
timeout_in_minutes: 100
101
depends_on:
102
- image-build-amd
.buildkite/test_areas/models_multimodal.yaml
+4
-4
@@ -15,7 +15,7 @@ steps:
15
- pytest -v -s models/multimodal/generation/test_ultravox.py -m core_model
16
mirror:
17
amd:
18
- device: mi300_1
18
+ device: mi325_1
19
depends_on:
20
- image-build-amd
21
@@ -33,7 +33,7 @@ steps:
33
- pytest -v -s models/multimodal/generation/test_vit_cudagraph.py -m core_model
34
mirror:
35
amd:
36
- device: mi300_1
36
+ device: mi325_1
37
depends_on:
38
- image-build-amd
39
@@ -50,7 +50,7 @@ steps:
50
- pytest -v -s models/multimodal/generation/test_qwen2_vl.py -m core_model
51
mirror:
52
amd:
53
- device: mi300_1
53
+ device: mi325_1
54
depends_on:
55
- image-build-amd
56
@@ -118,7 +118,7 @@ steps:
118
- pytest -v -s models/multimodal/test_mapping.py
119
mirror:
120
amd:
121
- device: mi300_1
121
+ device: mi325_1
122
depends_on:
123
- image-build-amd
124