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, 05:30 UTC
[Doc] improve test-pipeline.yaml documentation (#25305)
Signed-off-by: Huamin Li <3ericli@gmail.com> Co-authored-by: Lu Fang <30275821+houseroad@users.noreply.github.com>
Huamin Li committed
Sep 20, 2025 at 20:29 UTC
62b38dc8322f9f31fff7c680028b0b88306e1c39
1 file changed
+18
-14
.buildkite/test-pipeline.yaml
+18
-14
@@ -6,24 +6,28 @@
6
# to generate the final pipeline yaml file.
7
8
# Documentation
9
-# label(str): the name of the test. emoji allowed.
10
-# fast_check(bool): whether to run this on each commit on fastcheck pipeline.
11
-# torch_nightly(bool): whether to run this on vllm against torch nightly pipeline.
12
-# fast_check_only(bool): run this test on fastcheck pipeline only
13
-# optional(bool): never run this test by default (i.e. need to unblock manually) unless it's scheduled nightly run.
9
+# label(str): the name of the test. emojis allowed.
10
+# fast_check(bool): whether to run this on each commit on the fastcheck pipeline.
11
+# torch_nightly(bool): whether to run this on vllm against the torch nightly pipeline.
12
+# fast_check_only(bool): run this test on the fastcheck pipeline only
13
+# optional(bool): never run this test by default (i.e. need to unblock manually) unless it's a scheduled nightly run.
14
+# soft_fail(bool): allow this step to fail without failing the entire pipeline (useful for flaky or experimental tests).
15
# command(str): the single command to run for tests. incompatible with commands.
15
-# commands(list): the list of commands to run for test. incompatbile with command.
16
-# mirror_hardwares(list): the list of hardwares to run the test on as well. currently only supports [amd]
17
-# gpu(str): override the GPU selection for the test. default is on L4 GPUs. currently only supports a100
18
-# num_gpus(int): override the number of GPUs for the test. default to 1 GPU. currently support 2,4.
19
-# num_nodes(int): whether to simulate multi-node setup by launch multiple containers on one host,
20
-# in this case, commands must be specified. the first command runs on first host, the second
16
+# commands(list): the list of commands to run for the test. incompatible with command.
17
+# mirror_hardwares(list): the list of hardware to run the test on as well. currently only supports [amdexperimental]
18
+# gpu(str): override the GPU selection for the test. default is L4 GPUs. supports a100, b200, h200
19
+# num_gpus(int): override the number of GPUs for the test. defaults to 1 GPU. currently supports 2,4.
20
+# num_nodes(int): whether to simulate multi-node setup by launching multiple containers on one host,
21
+# in this case, commands must be specified. the first command runs on the first host, the second
22
# command runs on the second host.
22
-# working_dir(str): specify the place where command should execute, default to /vllm-workspace/tests
23
-# source_file_dependencies(list): the list of prefix to opt-in the test for, if empty, the test will always run.
23
+# timeout_in_minutes(int): sets a timeout for the step in minutes. if not specified, uses the default timeout.
24
+# parallelism(int): number of parallel jobs to run for this step. enables test sharding using $$BUILDKITE_PARALLEL_JOB
25
+# and $$BUILDKITE_PARALLEL_JOB_COUNT environment variables.
26
+# working_dir(str): specify the place where the command should execute, default to /vllm-workspace/tests
27
+# source_file_dependencies(list): the list of prefixes to opt-in the test for, if empty, the test will always run.
28
29
# When adding a test
26
-# - If the test belong to an existing group, add it there
30
+# - If the test belongs to an existing group, add it there
31
# - If the test is short, add to any existing step
32
# - If the test takes more than 10min, then it is okay to create a new step.
33
# Note that all steps execute in parallel.