EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

ACLgraph enable: Test cases revisions for all features (#3388)

Browse Source 
### What this PR does / why we need it?
This PR revise the test cases of various features on the warehouse which
add the enablement of aclgraph to the test cases.

### Does this PR introduce _any_ user-facing change?
no

### How was this patch tested?
ut

- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

Signed-off-by: lilinsiman <lilinsiman@gmail.com>
This commit is contained in:
lilinsiman
2025-10-17 17:15:19 +08:00
committed by GitHub
parent bf87606932
commit 1b424fb7f1
17 changed files with 34 additions and 117 deletions
Download Patch File Download Diff File Expand all files Collapse all files

81
tests/e2e/multicard/test_weight_loader.py
View File

@@ -33,47 +33,7 @@ DEVICE_NAME = torch_npu.npu.get_device_name(0)[:10]
@pytest.mark.parametrize("model", MOE_MODELS)
def test_external_launcher_eager(model):
script = script = "/usr/local/python3.11.13/bin/python3.11/__w/vllm-ascend/tests/examples/test_weight_loader.py"
env = os.environ.copy()
# TODO: Change to 2 when ci machine has 4 cards
cmd = [
sys.executable,
str(script),
"--model",
model,
"--tp-size",
"2",
"--proc-per-node",
"2",
"--trust-remote-code",
"--enforce-eager",
"--enable-expert-parallel",
"--enable-sleep-mode",
"--model-weight-gib",
"20",
]
print(f"Running subprocess: {' '.join(cmd)}")
proc = subprocess.run(
cmd,
env=env,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
timeout=600,
)
output = proc.stdout.decode()
print(output)
assert "TP RANKS: [0]" in output
assert "TP RANKS: [1]" in output
assert "Generated text:" in output
assert proc.returncode == 0
@pytest.mark.parametrize("model", MOE_MODELS)
def test_external_launcher_aclgraph(model):
def test_external_launcher(model):
script = "/usr/local/python3.11.13/bin/python3.11/__w/vllm-ascend/tests/examples/test_weight_loader.py"
env = os.environ.copy()
# TODO: Change to 2 when ci machine has 4 cards
@@ -147,42 +107,3 @@ def test_external_launcher_dense(model):
assert "TP RANKS: [1]" in output
assert "Generated text:" in output
assert proc.returncode == 0
@pytest.mark.parametrize("model", MODELS)
def test_external_launcher_dense_eager(model):
script = "/usr/local/python3.11.13/bin/python3.11/__w/vllm-ascend/tests/examples/test_weight_loader.py"
env = os.environ.copy()
# TODO: Change to 2 when ci machine has 4 cards
cmd = [
sys.executable,
str(script),
"--model",
model,
"--tp-size",
"2",
"--proc-per-node",
"2",
"--trust-remote-code",
"--enforce-eager",
"--enable-sleep-mode",
"--model-weight-gib",
"20",
]
print(f"Running subprocess: {' '.join(cmd)}")
proc = subprocess.run(
cmd,
env=env,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
timeout=600,
)
output = proc.stdout.decode()
print(output)
assert "TP RANKS: [0]" in output
assert "TP RANKS: [1]" in output
assert "Generated text:" in output
assert proc.returncode == 0