Update simd compatibility test (#27880)

Signed-off-by: yanliang567 <yanliang.qiao@zilliz.com>
pull/27927/head
yanliang567 2023-10-25 17:20:14 +08:00 committed by GitHub
parent de26231976
commit 8b5b33fce0
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 65 additions and 48 deletions

View File

@ -58,7 +58,7 @@ class MilvusOperator(object):
Return: custom resource object instance
"""
new_configs = self._update_configs(configs, template)
print(new_configs)
log.debug(new_configs)
namespace = new_configs['metadata'].get('namespace', 'default')
# apply custom resource object to deploy milvus
cus_res = CusResource(kind=self.plural, group=self.group,

View File

@ -1,29 +1,32 @@
apiVersion: milvus.io/v1alpha1
kind: MilvusCluster
apiVersion: milvus.io/v1beta1
kind: Milvus
metadata:
name: my-release
namespace: chaos-testing
labels:
app: milvus
spec:
config:
log:
level: debug
common:
simdType: avx
components: {}
dependencies:
etcd:
inCluster:
deletionPolicy: Delete
pvcDeletion: true
pulsar:
inCluster:
deletionPolicy: Delete
pvcDeletion: true
values:
proxy:
configData:
httpNumThreads: "100"
metrics:
podMonitor:
enabled: true
storage:
inCluster:
deletionPolicy: Delete
pvcDeletion: true
config:
log:
level: debug
components: {}
values:
metrics:
podMonitor:
enabled: true

View File

@ -24,11 +24,10 @@ namespace = 'chaos-testing'
def _install_milvus(simd):
release_name = f"mil-{simd.replace('_','-')}-" + cf.gen_digits_by_length(6)
cus_configs = {'spec.components.image': 'milvusdb/milvus:master-latest',
cus_configs = {'spec.components.image': 'harbor.milvus.io/milvus/milvus:master-latest',
'metadata.namespace': namespace,
'metadata.name': release_name,
'spec.components.proxy.serviceType': 'LoadBalancer',
'spec.config.knowhere.simdType': simd
'spec.config.common.simdType': simd
}
milvus_op = MilvusOperator()
log.info(f"install milvus with configs: {cus_configs}")
@ -52,17 +51,17 @@ class TestSimdCompatibility:
milvus_op.uninstall(self.release_name, namespace)
@pytest.mark.tags(CaseLabel.L3)
@pytest.mark.parametrize('simd_id', [i for i in range(len(supported_simd_types))])
def test_simd_compat_e2e(self, simd_id):
@pytest.mark.parametrize('simd', supported_simd_types)
def test_simd_compat_e2e(self, simd):
"""
steps
1. [test_milvus_install]: set up milvus with customized simd configured
2. [test_simd_compat_e2e]: verify milvus is working well
4. [test_milvus_cleanup]: delete milvus instances in teardown
"""
simd = supported_simd_types[simd_id]
log.info(f"start to install milvus with simd {simd}")
release_name, host, port = _install_milvus(simd)
time.sleep(10)
self.release_name = release_name
assert host is not None
conn = connections.connect("default", host=host, port=port)
@ -70,38 +69,55 @@ class TestSimdCompatibility:
mil = MilvusSys(alias="default")
log.info(f"milvus build version: {mil.build_version}")
log.info(f"milvus simdType: {mil.simd_type}")
assert str(mil.simd_type).lower() in [simd_type.lower() for simd_type in supported_simd_types[simd_id:]]
assert str(mil.simd_type).lower() == simd.lower()
log.info(f"start to e2e verification: {simd}")
# create
name = cf.gen_unique_str("compat")
prefix = "simd_"
name = cf.gen_unique_str(prefix)
t0 = time.time()
collection_w = ApiCollectionWrapper()
collection_w.init_collection(name=name,
schema=cf.gen_default_collection_schema(),
timeout=40)
collection_w.init_collection_wrap(name=name)
tt = time.time() - t0
assert collection_w.name == name
entities = collection_w.num_entities
log.info(f"assert create collection: {tt}, init_entities: {entities}")
# insert
data = cf.gen_default_list_data()
t0 = time.time()
_, res = collection_w.insert(data)
tt = time.time() - t0
log.info(f"assert insert: {tt}")
assert res
for _ in range(10):
data = cf.gen_default_list_data(nb=300)
t0 = time.time()
_, res = collection_w.insert(data)
tt = time.time() - t0
log.info(f"assert insert: {tt}")
assert res
# flush
t0 = time.time()
_, check_result = collection_w.flush(timeout=180)
assert check_result
assert collection_w.num_entities == len(data[0]) + entities
tt = time.time() - t0
entities = collection_w.num_entities
log.info(f"assert flush: {tt}, entities: {entities}")
# search
# index
index_params = {"index_type": "IVF_SQ8", "params": {"nlist": 64}, "metric_type": "L2"}
t0 = time.time()
index, _ = collection_w.create_index(field_name=ct.default_float_vec_field_name,
index_params=index_params,
index_name=cf.gen_unique_str())
index, _ = collection_w.create_index(field_name=ct.default_string_field_name,
index_params={},
index_name=cf.gen_unique_str())
tt = time.time() - t0
log.info(f"assert index: {tt}")
assert len(collection_w.indexes) == 2
# load
collection_w.load()
# search
search_vectors = cf.gen_vectors(1, ct.default_dim)
search_params = {"metric_type": "L2", "params": {"nprobe": 16}}
t0 = time.time()
@ -111,37 +127,35 @@ class TestSimdCompatibility:
tt = time.time() - t0
log.info(f"assert search: {tt}")
assert len(res_1) == 1
# release
collection_w.release()
# index
# insert
d = cf.gen_default_list_data()
collection_w.insert(d)
log.info(f"assert index entities: {collection_w.num_entities}")
_index_params = {"index_type": "IVF_SQ8", "params": {"nlist": 64}, "metric_type": "L2"}
t0 = time.time()
index, _ = collection_w.create_index(field_name=ct.default_float_vec_field_name,
index_params=_index_params,
name=cf.gen_unique_str())
tt = time.time() - t0
log.info(f"assert index: {tt}")
assert len(collection_w.indexes) == 1
# search
t0 = time.time()
collection_w.load()
tt = time.time() - t0
log.info(f"assert load: {tt}")
search_vectors = cf.gen_vectors(1, ct.default_dim)
nq = 5
topk = 5
search_vectors = cf.gen_vectors(nq, ct.default_dim)
t0 = time.time()
res_1, _ = collection_w.search(data=search_vectors,
anns_field=ct.default_float_vec_field_name,
param=search_params, limit=1)
res, _ = collection_w.search(data=search_vectors,
anns_field=ct.default_float_vec_field_name,
param=search_params, limit=topk)
tt = time.time() - t0
log.info(f"assert search: {tt}")
assert len(res) == nq
assert len(res[0]) <= topk
# query
term_expr = f'{ct.default_int64_field_name} in [1001,1201,4999,2999]'
term_expr = f'{ct.default_int64_field_name} in [1, 2, 3, 4]'
t0 = time.time()
res, _ = collection_w.query(term_expr)
tt = time.time() - t0
log.info(f"assert query result {len(res)}: {tt}")
assert len(res) >= 4