[CI] execute all piecewise compilation tests together (#24502)

Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>

.buildkite/test-pipeline.yaml

@@ -379,11 +379,7 @@ steps:
   - tests/compile
   commands:
   - pytest -v -s compile/test_basic_correctness.py
-  # these tests need to be separated, cannot combine
+  - pytest -v -s compile/piecewise/
-  - pytest -v -s compile/piecewise/test_simple.py
-  - pytest -v -s compile/piecewise/test_toy_llama.py
-  - pytest -v -s compile/piecewise/test_full_cudagraph.py
-  - pytest -v -s compile/piecewise/test_multiple_graphs.py
 
 - label: PyTorch Fullgraph Test # 20min
   timeout_in_minutes: 30

tests/compile/piecewise/test_multiple_graphs.py

@@ -4,9 +4,9 @@
 Test (piecewise) compilation with a simple model where multiple submodules
 are compiled and graph captured separately.
 """
+
 import torch
 from torch import nn
-from torch.library import Library
 
 from vllm.compilation.backends import set_model_tag
 from vllm.compilation.counter import compilation_counter
@@ -15,10 +15,9 @@ from vllm.compilation.decorators import (ignore_torch_compile,
 from vllm.config import (CompilationConfig, CompilationLevel, CUDAGraphMode,
                          VllmConfig, set_current_vllm_config)
 from vllm.forward_context import BatchDescriptor, set_forward_context
-from vllm.utils import direct_register_custom_op
 
-# create a library to hold the custom op
+# This import automatically registers `torch.ops.silly.attention`
-silly_lib = Library("silly", "FRAGMENT")  # noqa
+from .. import silly_attention  # noqa: F401
 
 BATCH_SIZE = 32
 MLP_SIZE = 128
@@ -26,27 +25,6 @@ HIDDEN_SIZE = 1024
 RANDOM_SEED = 0
 
 
-def silly_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                    out: torch.Tensor) -> None:
-    out.copy_(q)
-    out += k
-    out += v
-
-
-def silly_attention_fake(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                         out: torch.Tensor) -> None:
-    return
-
-
-direct_register_custom_op(
-    op_name="attention",
-    op_func=silly_attention,
-    mutates_args=["out"],
-    fake_impl=silly_attention_fake,
-    target_lib=silly_lib,
-)
-
-
 @support_torch_compile
 class ParentModel(nn.Module):
 

tests/compile/piecewise/test_simple.py

@@ -4,10 +4,10 @@
 Test the piecewise compilation with a simple model so that we
 can exactly calculate the expected output and side effects.
 """
+
 import pytest
 import torch
 from torch import nn
-from torch.library import Library
 
 from vllm.compilation.counter import compilation_counter
 from vllm.compilation.decorators import support_torch_compile
@@ -15,35 +15,9 @@ from vllm.config import (CompilationConfig, CompilationLevel, CUDAGraphMode,
                          VllmConfig, set_current_vllm_config)
 from vllm.envs import VLLM_USE_V1
 from vllm.forward_context import BatchDescriptor, set_forward_context
-from vllm.utils import direct_register_custom_op
 
-global_counter = 0
+# This import automatically registers `torch.ops.silly.attention`
-
+from ..silly_attention import get_global_counter, reset_global_counter
-# create a library to hold the custom op
-silly_lib = Library("silly", "FRAGMENT")  # noqa
-
-
-def silly_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                    out: torch.Tensor) -> None:
-    global global_counter
-    global_counter += 1
-    print(f"{global_counter=}")
-    out.copy_(q)
-    out[0] += 1
-
-
-def silly_attention_fake(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                         out: torch.Tensor) -> None:
-    return
-
-
-direct_register_custom_op(
-    op_name="attention",
-    op_func=silly_attention,
-    mutates_args=["out"],
-    fake_impl=silly_attention_fake,
-    target_lib=silly_lib,
-)
 
 
 @support_torch_compile
@@ -59,8 +33,7 @@ class SillyModel(nn.Module):
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         """
         Overall effect:
-        x += 1
+        x = 3 * x + 19
-        x[0] += 2
         global_counter += 2
         """
         x = x + 1
@@ -78,6 +51,7 @@ class SillyModel(nn.Module):
 
 
 @pytest.mark.parametrize("use_inductor", [True, False])
+@torch.inference_mode()
 def test_simple_piecewise_compile(use_inductor):
     assert VLLM_USE_V1
 
@@ -121,13 +95,12 @@ def test_simple_piecewise_compile(use_inductor):
             model(torch.randn(1).cuda())
 
         input = torch.zeros(2).cuda()
-        global global_counter
+        reset_global_counter()
-        global_counter = 0
         with set_forward_context(
                 None,
                 vllm_config=vllm_config,
                 cudagraph_runtime_mode=CUDAGraphMode.PIECEWISE,
                 batch_descriptor=BatchDescriptor(num_tokens=2, )):
             output = model(input)
-        assert global_counter == 2
+        assert get_global_counter() == 2
-        assert torch.allclose(output.cpu(), torch.tensor([3., 1.]))
+        assert torch.allclose(output.cpu(), torch.tensor([19.0, 19.0]))

tests/compile/piecewise/test_toy_llama.py

@@ -14,38 +14,15 @@ from typing import Any, Optional
 import pytest
 import torch
 from torch import nn
-from torch.library import Library
 
 from vllm.compilation.counter import compilation_counter
 from vllm.compilation.decorators import support_torch_compile
 from vllm.config import (CompilationConfig, CompilationLevel, CUDAGraphMode,
                          VllmConfig, set_current_vllm_config)
 from vllm.forward_context import BatchDescriptor, set_forward_context
-from vllm.utils import direct_register_custom_op
 
-# create a library to hold the custom op
+# This import automatically registers `torch.ops.silly.attention`
-silly_lib = Library("silly", "FRAGMENT")  # noqa
+from .. import silly_attention  # noqa: F401
-
-
-def silly_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                    out: torch.Tensor) -> None:
-    out.copy_(q)
-    out += k
-    out += v
-
-
-def silly_attention_fake(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                         out: torch.Tensor) -> None:
-    return
-
-
-direct_register_custom_op(
-    op_name="attention",
-    op_func=silly_attention,
-    mutates_args=["out"],
-    fake_impl=silly_attention_fake,
-    target_lib=silly_lib,
-)
 
 
 @dataclass

tests/compile/silly_attention.py

@@ -0,0 +1,63 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+"""
+Shared PyTorch custom silly attention for compilation tests.
+Centralizes custom operation definitions to avoid duplicate registrations.
+"""
+
+import torch
+from torch.library import Library
+
+from vllm.utils import direct_register_custom_op
+
+# Shared library for all compilation test operations
+# Using "silly" namespace to match existing test expectations
+# import this file will automatically register
+# torch ops for testing (like silly.attention)
+silly_lib = Library("silly", "FRAGMENT")
+
+# Global counter that counts the number of times attention is invoked
+_global_counter = 0
+
+
+def get_global_counter():
+    """Get the current global counter value"""
+    return _global_counter
+
+
+def reset_global_counter():
+    """Reset the global counter to 0"""
+    global _global_counter
+    _global_counter = 0
+
+
+def silly_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
+                    out: torch.Tensor) -> None:
+    """
+    Unified attention implementation that depends on
+    all inputs and affects the output.
+    Always increments a global counter that tests can use or ignore.
+    """
+    global _global_counter
+
+    # Always increment the global counter
+    _global_counter += 1
+
+    # Unified implementation that depends on all inputs
+    out.copy_(q + k + v)
+
+
+def silly_attention_fake(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
+                         out: torch.Tensor) -> None:
+    """Fake implementation for testing"""
+    return
+
+
+# Register the unified attention operation
+direct_register_custom_op(
+    op_name="attention",
+    op_func=silly_attention,
+    mutates_args=["out"],
+    fake_impl=silly_attention_fake,
+    target_lib=silly_lib,
+)

tests/compile/test_decorator.py

@@ -2,7 +2,6 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 import torch
 from torch import nn
-from torch.library import Library
 
 from vllm.compilation.counter import compilation_counter
 from vllm.compilation.decorators import (ignore_torch_compile,
@@ -10,36 +9,14 @@ from vllm.compilation.decorators import (ignore_torch_compile,
 from vllm.config import (CacheConfig, CompilationConfig, CompilationLevel,
                          CUDAGraphMode, VllmConfig, set_current_vllm_config)
 from vllm.forward_context import BatchDescriptor, set_forward_context
-from vllm.utils import direct_register_custom_op
 
-# create a library to hold the custom op
+# This import automatically registers `torch.ops.silly.attention`
-silly_lib = Library("silly", "FRAGMENT")  # noqa
+from . import silly_attention  # noqa: F401
 
 BATCH_SIZE = 32
 MLP_SIZE = 128
 
 
-def silly_attention(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                    out: torch.Tensor) -> None:
-    out.copy_(q)
-    out += k
-    out += v
-
-
-def silly_attention_fake(q: torch.Tensor, k: torch.Tensor, v: torch.Tensor,
-                         out: torch.Tensor) -> None:
-    return
-
-
-direct_register_custom_op(
-    op_name="attention",
-    op_func=silly_attention,
-    mutates_args=["out"],
-    fake_impl=silly_attention_fake,
-    target_lib=silly_lib,
-)
-
-
 @torch.inference_mode
 def run_model(vllm_config: VllmConfig, model: nn.Module,
               cudagraph_runtime_mode: CUDAGraphMode):
@@ -151,7 +128,7 @@ def test_ignore_torch_compile_decorator():
         run_model(vllm_config, mod_C, cudagraph_runtime_mode)
 
 
-# Only enable torch.compile if
+# Only enable torch.compile if
 # vllm_config.cache_config.kv_sharing_fast_prefill=True
 @support_torch_compile(enable_if=lambda vllm_config: vllm_config.cache_config.
                        kv_sharing_fast_prefill)
@@ -173,7 +150,7 @@ class B(nn.Module):
         return x
 
 
-# Only enable torch.compile if
+# Only enable torch.compile if
 # vllm_config.cache_config.kv_sharing_fast_prefill=False
 @support_torch_compile(enable_if=lambda vllm_config: not vllm_config.
                        cache_config.kv_sharing_fast_prefill)