EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[HICache]: Refactor HiCache CI (#11011)

Browse Source 
Co-authored-by: pansicheng <sicheng.pan.chn@gmail.com>
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
This commit is contained in:
hzh0425
2025-10-04 08:51:56 +08:00
committed by GitHub
parent c61b9a1d01
commit c70e58e837
4 changed files with 81 additions and 193 deletions
Download Patch File Download Diff File Expand all files Collapse all files

53
test/srt/hicache/test_hicache_storage_3fs_backend.py
View File

@@ -62,19 +62,7 @@ class TestHf3fsBackendLayerFirstLayout(
server_args, env_vars = super()._get_additional_server_args_and_env() server_args, env_vars = super()._get_additional_server_args_and_env()
server_args["--hicache-mem-layout"] = "layer_first" server_args["--hicache-mem-layout"] = "layer_first"
server_args["--hicache-io-backend"] = "direct" server_args["--hicache-io-backend"] = "direct"
return server_args, env_vars server_args["--tp-size"] = 2
class TestHf3fsBackendPageFirstLayout(
HiCacheStorage3FSBackendBaseMixin, CustomTestCase
):
"""Page first layout tests for HiCache-Hf3fs backend"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args, env_vars = super()._get_additional_server_args_and_env()
server_args["--hicache-mem-layout"] = "page_first"
return server_args, env_vars return server_args, env_vars
@@ -91,44 +79,9 @@ class TestHf3fsBackendAccuracy(HiCacheStorage3FSBackendBaseMixin, CustomTestCase
def test_eval_accuracy(self): def test_eval_accuracy(self):
"""Test eval accuracy with cache persistence across cache flushes""" """Test eval accuracy with cache persistence across cache flushes"""
print("\n=== Testing Eval Accuracy with Cache Persistence ===") from test_hicache_storage_file_backend import run_eval_accuracy_test
# First evaluation - populate cache run_eval_accuracy_test(self)
print("Phase 1: Running initial GSM8K evaluation to populate cache...")
args_initial = SimpleNamespace(
num_shots=5,
data_path=None,
num_questions=50,
max_new_tokens=512,
parallel=10,
host=f"http://{self.base_host}",
port=int(self.base_port),
)
metrics_initial = run_eval_few_shot_gsm8k(args_initial)
# Flush cache to force remote storage access
print("Phase 2: Flushing device cache...")
self.assertTrue(self.flush_cache(), "Cache flush should succeed")
time.sleep(2)
# Second evaluation - should use remote cache
print("Phase 3: Running second GSM8K evaluation using remote cache...")
metrics_cached = run_eval_few_shot_gsm8k(args_initial)
# Verify accuracy consistency
accuracy_diff = abs(metrics_initial["accuracy"] - metrics_cached["accuracy"])
print(f"Accuracy difference: {accuracy_diff:.4f}")
# Assertions
self.assertGreater(
metrics_initial["accuracy"], 0.6, "Initial accuracy should be reasonable"
)
self.assertGreater(
metrics_cached["accuracy"], 0.6, "Cached accuracy should be reasonable"
)
self.assertLess(
accuracy_diff, 0.05, "Accuracy should be consistent between cache states"
)
if __name__ == "__main__": if __name__ == "__main__":

157
test/srt/hicache/test_hicache_storage_file_backend.py
View File

@@ -24,6 +24,7 @@ from sglang.test.test_utils import (
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH, DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
DEFAULT_URL_FOR_TEST, DEFAULT_URL_FOR_TEST,
CustomTestCase, CustomTestCase,
is_in_ci,
popen_launch_server, popen_launch_server,
) )
@@ -89,6 +90,7 @@ class HiCacheStorageBaseMixin:
"""Launch server with HiCache enabled""" """Launch server with HiCache enabled"""
additional_server_args, env_vars = cls._get_additional_server_args_and_env() additional_server_args, env_vars = cls._get_additional_server_args_and_env()
env_vars["SGLANG_ENABLE_DETERMINISTIC_INFERENCE"] = "1"
server_args = cls._get_base_server_args() server_args = cls._get_base_server_args()
if additional_server_args: if additional_server_args:
server_args.update(additional_server_args) server_args.update(additional_server_args)
@@ -215,42 +217,7 @@ class HiCacheStorageBaseMixin:
) )
class TestHiCacheStorageTP(HiCacheStorageBaseMixin, CustomTestCase): @unittest.skipIf(is_in_ci(), "To reduce the CI execution time.")
"""Multi-TP tests for HiCache Storage functionality"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args = {"--tp-size": 2}
return server_args, {}
class TestHiCacheStorageLayerFirstDirectIO(HiCacheStorageBaseMixin, CustomTestCase):
"""Layer first direct tests for HiCache Storage functionality"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args = {
"--hicache-mem-layout": "layer_first",
"--hicache-io-backend": "direct",
}
return server_args, {}
class TestHiCacheStoragePageFirstDirectIO(HiCacheStorageBaseMixin, CustomTestCase):
"""Page first direct tests for HiCache Storage functionality"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args = {
"--hicache-mem-layout": "page_first_direct",
"--hicache-io-backend": "direct",
}
return server_args, {}
class TestHiCacheStoragePageFirstLayout(HiCacheStorageBaseMixin, CustomTestCase): class TestHiCacheStoragePageFirstLayout(HiCacheStorageBaseMixin, CustomTestCase):
"""Page first layout tests for HiCache Storage functionality""" """Page first layout tests for HiCache Storage functionality"""
@@ -261,6 +228,7 @@ class TestHiCacheStoragePageFirstLayout(HiCacheStorageBaseMixin, CustomTestCase)
return server_args, {} return server_args, {}
@unittest.skipIf(is_in_ci(), "To reduce the CI execution time.")
class TestHiCacheStorageMLA(HiCacheStorageBaseMixin, CustomTestCase): class TestHiCacheStorageMLA(HiCacheStorageBaseMixin, CustomTestCase):
"""MLA Model tests for HiCache Storage functionality""" """MLA Model tests for HiCache Storage functionality"""
@@ -276,71 +244,84 @@ class TestHiCacheStorageMLA(HiCacheStorageBaseMixin, CustomTestCase):
return server_args, {} return server_args, {}
class TestHiCacheStoragePageFirstDirectIO(HiCacheStorageBaseMixin, CustomTestCase):
"""Page first direct tests for HiCache Storage functionality"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args = {
"--hicache-mem-layout": "page_first_direct",
"--hicache-io-backend": "direct",
"--tp-size": 2,
}
return server_args, {}
class TestHiCacheStorageAccuracy(HiCacheStorageBaseMixin, CustomTestCase): class TestHiCacheStorageAccuracy(HiCacheStorageBaseMixin, CustomTestCase):
"""Accuracy tests for HiCache Storage functionality""" """Accuracy tests for HiCache Storage functionality"""
@classmethod @classmethod
def _get_additional_server_args_and_env(cls): def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses""" """Get additional server arguments specific to configuration - override in subclasses"""
server_args = {"--tp-size": 2, "--hicache-ratio": 1.5} server_args = {
"--tp-size": 2,
"--hicache-ratio": 1.5,
}
return server_args, {} return server_args, {}
def test_eval_accuracy(self): def test_eval_accuracy(self):
"""Test eval accuracy with cache persistence across cache flushes""" """Test eval accuracy with cache persistence across cache flushes"""
print("\n=== Testing Eval Accuracy with Cache Persistence ===") run_eval_accuracy_test(self)
# First evaluation - populate cache
print("Phase 1: Running initial GSM8K evaluation to populate cache...")
args_initial = SimpleNamespace(
num_shots=5,
data_path=None,
num_questions=50,
max_new_tokens=512,
parallel=10,
host=f"http://{self.base_host}",
port=int(self.base_port),
)
metrics_initial = run_eval_few_shot_gsm8k(args_initial)
# Flush cache to force remote storage access
print("Phase 2: Flushing device cache...")
self.assertTrue(self.flush_cache(), "Cache flush should succeed")
time.sleep(2)
# Second evaluation - should use remote cache
print("Phase 3: Running second GSM8K evaluation using remote cache...")
metrics_cached = run_eval_few_shot_gsm8k(args_initial)
# Verify accuracy consistency
accuracy_diff = abs(metrics_initial["accuracy"] - metrics_cached["accuracy"])
print(f"Accuracy difference: {accuracy_diff:.4f}")
# Assertions
self.assertGreater(
metrics_initial["accuracy"], 0.6, "Initial accuracy should be reasonable"
)
self.assertGreater(
metrics_cached["accuracy"], 0.6, "Cached accuracy should be reasonable"
)
self.assertLess(
accuracy_diff, 0.05, "Accuracy should be consistent between cache states"
)
# TODO: Add other backends tests（3fs/mooncake） def run_eval_accuracy_test(test_instance, accuracy_threshold: float = 0.03):
# class TestHiCacheStorageMooncakeBackend(HiCacheStorageBaseTest): """Generic eval accuracy test with configurable accuracy threshold
# """Mooncake backend tests for HiCache Storage functionality"""
# @classmethod Args:
# def _get_additional_server_args_and_env(cls): test_instance: The test class instance that provides base_host, base_port, flush_cache, and assert methods
# """Get additional server arguments specific to configuration - override in subclasses""" """
# server_args = ["--hicache-storage-backend", "mooncake"] print("\n=== Testing Eval Accuracy with Cache Persistence ===")
# env = {
# "MOONCAKE_TE_META_DATA_SERVER": "http://127.0.0.1:8080/metadata", # First evaluation - populate cache
# "MOONCAKE_MASTER": "127.0.0.1:50051" print("Phase 1: Running initial GSM8K evaluation to populate cache...")
# xxxxx args_initial = SimpleNamespace(
# } num_shots=5,
# return server_args, {} data_path=None,
num_questions=50,
max_new_tokens=512,
parallel=10,
host=f"http://{test_instance.base_host}",
port=int(test_instance.base_port),
)
metrics_initial = run_eval_few_shot_gsm8k(args_initial)
# Flush cache to force remote storage access
print("Phase 2: Flushing device cache...")
test_instance.assertTrue(test_instance.flush_cache(), "Cache flush should succeed")
time.sleep(2)
# Second evaluation - should use remote cache
print("Phase 3: Running second GSM8K evaluation using remote cache...")
metrics_cached = run_eval_few_shot_gsm8k(args_initial)
# Verify accuracy consistency
accuracy_diff = abs(metrics_initial["accuracy"] - metrics_cached["accuracy"])
print(f"Accuracy difference: {accuracy_diff:.4f}")
# Assertions
test_instance.assertGreater(
metrics_initial["accuracy"], 0.6, "Initial accuracy should be reasonable"
)
test_instance.assertGreater(
metrics_cached["accuracy"], 0.6, "Cached accuracy should be reasonable"
)
test_instance.assertLess(
accuracy_diff,
accuracy_threshold,
"Accuracy should be consistent between cache states",
)
if __name__ == "__main__": if __name__ == "__main__":

58
test/srt/hicache/test_hicache_storage_mooncake_backend.py
View File

@@ -19,6 +19,7 @@ from sglang.test.test_utils import (
DEFAULT_MLA_MODEL_NAME_FOR_TEST, DEFAULT_MLA_MODEL_NAME_FOR_TEST,
CustomTestCase, CustomTestCase,
find_available_port, find_available_port,
is_in_ci,
) )
@@ -226,6 +227,7 @@ class TestMooncakeBackendLayerFirstLayout(
''' '''
@unittest.skipIf(is_in_ci(), "To reduce the CI execution time.")
class TestMooncakeBackendPageFirstLayout( class TestMooncakeBackendPageFirstLayout(
HiCacheStorageMooncakeBackendBaseMixin, CustomTestCase HiCacheStorageMooncakeBackendBaseMixin, CustomTestCase
): ):
@@ -236,21 +238,6 @@ class TestMooncakeBackendPageFirstLayout(
"""Get additional server arguments specific to configuration - override in subclasses""" """Get additional server arguments specific to configuration - override in subclasses"""
server_args, env_vars = super()._get_additional_server_args_and_env() server_args, env_vars = super()._get_additional_server_args_and_env()
server_args["--hicache-mem-layout"] = "page_first" server_args["--hicache-mem-layout"] = "page_first"
server_args["--hicache-io-backend"] = "kernel"
return server_args, env_vars
class TestMooncakeBackendPageFirstDirectLayout(
HiCacheStorageMooncakeBackendBaseMixin, CustomTestCase
):
"""Page first layout tests for HiCache-Mooncake backend"""
@classmethod
def _get_additional_server_args_and_env(cls):
"""Get additional server arguments specific to configuration - override in subclasses"""
server_args, env_vars = super()._get_additional_server_args_and_env()
server_args["--hicache-mem-layout"] = "page_first_direct"
server_args["--hicache-io-backend"] = "direct"
return server_args, env_vars return server_args, env_vars
@@ -284,48 +271,15 @@ class TestMooncakeBackendAccuracy(
server_args, env_vars = super()._get_additional_server_args_and_env() server_args, env_vars = super()._get_additional_server_args_and_env()
server_args["--hicache-ratio"] = 1.5 server_args["--hicache-ratio"] = 1.5
server_args["--tp-size"] = 2 server_args["--tp-size"] = 2
server_args["--hicache-mem-layout"] = "page_first_direct"
server_args["--hicache-io-backend"] = "direct"
return server_args, env_vars return server_args, env_vars
def test_eval_accuracy(self): def test_eval_accuracy(self):
"""Test eval accuracy with cache persistence across cache flushes""" """Test eval accuracy with cache persistence across cache flushes"""
print("\n=== Testing Eval Accuracy with Cache Persistence ===") from test_hicache_storage_file_backend import run_eval_accuracy_test
# First evaluation - populate cache run_eval_accuracy_test(self)
print("Phase 1: Running initial GSM8K evaluation to populate cache...")
args_initial = SimpleNamespace(
num_shots=5,
data_path=None,
num_questions=50,
max_new_tokens=512,
parallel=10,
host=f"http://{self.base_host}",
port=int(self.base_port),
)
metrics_initial = run_eval_few_shot_gsm8k(args_initial)
# Flush cache to force remote storage access
print("Phase 2: Flushing device cache...")
self.assertTrue(self.flush_cache(), "Cache flush should succeed")
time.sleep(2)
# Second evaluation - should use remote cache
print("Phase 3: Running second GSM8K evaluation using remote cache...")
metrics_cached = run_eval_few_shot_gsm8k(args_initial)
# Verify accuracy consistency
accuracy_diff = abs(metrics_initial["accuracy"] - metrics_cached["accuracy"])
print(f"Accuracy difference: {accuracy_diff:.4f}")
# Assertions
self.assertGreater(
metrics_initial["accuracy"], 0.6, "Initial accuracy should be reasonable"
)
self.assertGreater(
metrics_cached["accuracy"], 0.6, "Cached accuracy should be reasonable"
)
self.assertLess(
accuracy_diff, 0.05, "Accuracy should be consistent between cache states"
)
if __name__ == "__main__": if __name__ == "__main__":

6
test/srt/run_suite.py
View File

@@ -132,8 +132,8 @@ suites = {
TestFile("test_load_weights_from_remote_instance.py", 72), TestFile("test_load_weights_from_remote_instance.py", 72),
TestFile("test_patch_torch.py", 19), TestFile("test_patch_torch.py", 19),
TestFile("test_release_memory_occupation.py", 257), TestFile("test_release_memory_occupation.py", 257),
TestFile("hicache/test_hicache_storage_file_backend.py", 400), TestFile("hicache/test_hicache_storage_file_backend.py", 200),
TestFile("hicache/test_hicache_storage_3fs_backend.py", 400), TestFile("hicache/test_hicache_storage_3fs_backend.py", 200),
], ],
"per-commit-4-gpu": [ "per-commit-4-gpu": [
TestFile("test_gpt_oss_4gpu.py", 600), TestFile("test_gpt_oss_4gpu.py", 600),
@@ -144,7 +144,7 @@ suites = {
TestFile("test_multi_instance_release_memory_occupation.py", 64), TestFile("test_multi_instance_release_memory_occupation.py", 64),
], ],
"per-commit-8-gpu": [ "per-commit-8-gpu": [
TestFile("hicache/test_hicache_storage_mooncake_backend.py", 800), TestFile("hicache/test_hicache_storage_mooncake_backend.py", 400),
TestFile("lora/test_lora_llama4.py", 600), TestFile("lora/test_lora_llama4.py", 600),
TestFile("test_disaggregation.py", 499), TestFile("test_disaggregation.py", 499),
TestFile("test_disaggregation_dp_attention.py", 155), TestFile("test_disaggregation_dp_attention.py", 155),