Skip to content

[TPU] Update dynamo dump file name in compilation test #19108

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 9 commits into from
Jun 4, 2025
Merged

[TPU] Update dynamo dump file name in compilation test #19108

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
5a9798a
fix test
lsy323 Jun 3, 2025
7fd4309
remove breakpoint
lsy323 Jun 3, 2025
559a206
add comment
lsy323 Jun 3, 2025
ba9e47d
add comment
lsy323 Jun 3, 2025
3973023
skip hanging tests
lsy323 Jun 3, 2025
705c99a
skip tests for ci disk size
lsy323 Jun 4, 2025
5c8b75e
fix quotes
lsy323 Jun 4, 2025
6f1aefe
disable some tests that failed from the begining
lsy323 Jun 4, 2025
f7f3a26
merge with main
lsy323 Jun 4, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .buildkite/scripts/hardware_ci/run-tpu-v1-test.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -150,7 +150,7 @@ run_and_track_test 9 "test_multimodal.py" \
run_and_track_test 10 "test_pallas.py" \
"python3 -m pytest -s -v /workspace/vllm/tests/v1/tpu/test_pallas.py"
run_and_track_test 11 "test_struct_output_generate.py" \
"python3 -m pytest -s -v /workspace/vllm/tests/v1/entrypoints/llm/test_struct_output_generate.py -k 'not test_structured_output_with_reasoning_matrices'"
"python3 -m pytest -s -v /workspace/vllm/tests/v1/entrypoints/llm/test_struct_output_generate.py -k \"not test_structured_output_with_reasoning_matrices\""
run_and_track_test 12 "test_moe_pallas.py" \
"python3 -m pytest -s -v /workspace/vllm/tests/tpu/test_moe_pallas.py"
run_and_track_test 13 "test_lora.py" \
Expand Down
5 changes: 3 additions & 2 deletions tests/tpu/test_compilation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,9 +64,10 @@ def extract_compiled_index(s):
numbers = [int(part) for part in parts if part.isdigit()]
return numbers[0]

# Check all the compilations are as expected
# Check all the compilations are as expected. The dump files include the
# captured graph for the forward function of the nn.Module.
compiled_fns = sorted(glob.glob(
os.path.join(temp_dir, "__compiled_fn*Captured*.py")),
os.path.join(temp_dir, "__compiled_fn*Forward_graph*.py")),
Comment on lines 69 to +70
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

Consider adding a comment explaining why __compiled_fn*Forward_graph*.py is now being checked instead of __compiled_fn*Captured*.py. This will help future developers understand the change and its context. It would be useful to add a brief explanation of what the Forward_graph represents.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks gemini, added a comment

key=lambda s: extract_compiled_index(s))

for i, compiled_fn in enumerate(compiled_fns):
Expand Down
25 changes: 15 additions & 10 deletions tests/v1/tpu/worker/test_tpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -370,6 +370,7 @@ def test_get_req_paddings():
assert _get_req_paddings(8, 36) == [8, 16, 32, 36]


@pytest.mark.skip(reason="Test is broken on TPU when it's added.")
def test_init_kv_cache_with_kv_sharing_invalid_target_layer_order():
layer_0 = "model.layers.0.self_attn.attn"
layer_1 = "model.layers.1.self_attn.attn"
Expand All @@ -381,15 +382,15 @@ def test_init_kv_cache_with_kv_sharing_invalid_target_layer_order():
layer_0:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_0,
kv_sharing_target_layer_name=layer_1,
),
layer_1:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_1,
)
Expand All @@ -398,6 +399,7 @@ def test_init_kv_cache_with_kv_sharing_invalid_target_layer_order():
assert fwd_context is not None


@pytest.mark.skip(reason="Test is broken on TPU when it's added.")
def test_init_kv_cache_with_kv_sharing_target_layer_not_exist():
layer_0 = "model.layers.0.self_attn.attn"
layer_1 = "model.layers.1.self_attn.attn"
Expand All @@ -408,14 +410,14 @@ def test_init_kv_cache_with_kv_sharing_target_layer_not_exist():
layer_0:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_0,
),
layer_1:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_1,
# invalid layer: cross_attn.atn doesn't exist!
Expand All @@ -426,6 +428,7 @@ def test_init_kv_cache_with_kv_sharing_target_layer_not_exist():
assert fwd_context is not None


@pytest.mark.skip(reason="Test is broken on TPU when it's added.")
def test_init_kv_cache_with_kv_sharing_target_same_as_current():
layer_0 = "model.layers.0.self_attn.attn"
layer_1 = "model.layers.1.self_attn.attn"
Expand All @@ -437,14 +440,14 @@ def test_init_kv_cache_with_kv_sharing_target_same_as_current():
layer_0:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_0,
),
layer_1:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_1,
kv_sharing_target_layer_name=layer_1,
Expand All @@ -454,6 +457,7 @@ def test_init_kv_cache_with_kv_sharing_target_same_as_current():
assert fwd_context is not None


@pytest.mark.skip(reason="Test is broken on TPU when it's added.")
def test_init_kv_cache_without_kv_sharing(model_runner):
layer_0 = "model.layers.0.self_attn.attn"
layer_1 = "model.layers.1.self_attn.attn"
Expand All @@ -463,14 +467,14 @@ def test_init_kv_cache_without_kv_sharing(model_runner):
layer_0:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_0,
),
layer_1:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_1,
)
Expand Down Expand Up @@ -520,6 +524,7 @@ def test_init_kv_cache_without_kv_sharing(model_runner):
assert kv_cache_config.kv_cache_groups[0].layer_names[1] == layer_1


@pytest.mark.skip(reason="Test is broken on TPU when it's added.")
def test_init_kv_cache_with_kv_sharing_valid(model_runner):
layer_0 = "model.layers.0.self_attn.attn"
layer_1 = "model.layers.1.self_attn.attn"
Expand All @@ -529,14 +534,14 @@ def test_init_kv_cache_with_kv_sharing_valid(model_runner):
layer_0:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_0,
),
layer_1:
Attention(
num_heads=8,
head_size=64,
head_size=128,
scale=1.0,
prefix=layer_1,
kv_sharing_target_layer_name="model.layers.0.self_attn.attn",
Expand Down