Benchmark updates/fixes

conda/environments/all_cuda-125_arch-x86_64.yaml

@@ -88,6 +88,7 @@ dependencies:
 - pydantic
 - pylibcudf=24.10
 - pylint=3.0.3
+- pynvml=11.4
 - pypdf=3.17.4
 - pypdfium2=4.30
 - pytest-asyncio

conda/environments/dev_cuda-125_arch-x86_64.yaml

@@ -73,6 +73,7 @@ dependencies:
 - pydantic
 - pylibcudf=24.10
 - pylint=3.0.3
+- pynvml=11.4
 - pypdfium2=4.30
 - pytest-asyncio
 - pytest-benchmark=4.0

dependencies.yaml

@@ -326,6 +326,7 @@ dependencies:
           - include-what-you-use=0.20
           - isort
           - pylint=3.0.3
+          - pynvml=11.4
           - vale=3.7
           - vale-styles-microsoft
           - vale-styles-write-good

tests/benchmarks/README.md

@@ -78,7 +78,6 @@ Fetch input data for benchmarks:
 ./scripts/fetch_data.py fetch validation
 ```
 
-
 ### Run E2E Benchmarks
 
 Benchmarks are run using `pytest-benchmark`. By default, there are five rounds of measurement. For each round, there will be one iteration of each workflow. Measurements are taken for each round. Final results such as `min`, `max` and `mean` times will be based on these measurements.
@@ -124,7 +123,6 @@ The `test_bench_e2e_pipelines.py` script contains several benchmarks within it.
 - `test_sid_nlp_e2e`
 - `test_abp_fil_e2e`
 - `test_phishing_nlp_e2e`
-- `test_cloudtrail_ae_e2e`
 
 For example, to run E2E benchmarks on the SID NLP workflow:
 ```bash
@@ -138,11 +136,10 @@ pytest -s --run_benchmark --benchmark-enable --benchmark-warmup=on --benchmark-w
 
 The console output should look like this:
 ```
---------------------------------------------------------------------------------- benchmark: 4 tests --------------------------------------------------------------------------------
+--------------------------------------------------------------------------------- benchmark: 3 tests --------------------------------------------------------------------------------
 Name (time in s)              Min               Max              Mean            StdDev            Median               IQR            Outliers     OPS            Rounds  Iterations
 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 test_sid_nlp_e2e           1.8907 (1.0)      1.9817 (1.0)      1.9400 (1.0)      0.0325 (2.12)     1.9438 (1.0)      0.0297 (1.21)          2;0  0.5155 (1.0)           5           1
-test_cloudtrail_ae_e2e     3.3403 (1.77)     3.3769 (1.70)     3.3626 (1.73)     0.0153 (1.0)      3.3668 (1.73)     0.0245 (1.0)           1;0  0.2974 (0.58)          5           1
 test_abp_fil_e2e           5.1271 (2.71)     5.3044 (2.68)     5.2083 (2.68)     0.0856 (5.59)     5.1862 (2.67)     0.1653 (6.75)          1;0  0.1920 (0.37)          5           1
 test_phishing_nlp_e2e      5.6629 (3.00)     6.0987 (3.08)     5.8835 (3.03)     0.1697 (11.08)    5.8988 (3.03)     0.2584 (10.55)         2;0  0.1700 (0.33)          5           1
 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

tests/benchmarks/e2e_test_configs.json

@@ -1,5 +1,5 @@
 {
-    "triton_server_url": "localhost:8001",
+    "triton_server_url": "localhost:8000",
     "test_sid_nlp_e2e": {
         "file_path": "../../models/datasets/validation-data/sid-validation-data.csv",
         "repeat": 10,

tests/benchmarks/test_bench_completion_pipeline.py

@@ -14,6 +14,7 @@
 # limitations under the License.
 
 import collections.abc
+import os
 import typing
 
 import pytest
@@ -82,4 +83,6 @@ def test_completion_pipe(benchmark: collections.abc.Callable[[collections.abc.Ca
                          config: Config,
                          dataset: DatasetManager,
                          llm_service_cls: type[LLMService]):
+    if llm_service_cls == OpenAIChatService:
+        os.environ.update({"OPENAI_API_KEY": "test_api_key"})
     benchmark(_run_pipeline, config, llm_service_cls, source_df=dataset["countries.csv"])

tests/benchmarks/test_bench_e2e_pipelines.py

@@ -35,7 +35,7 @@
 from morpheus.stages.preprocess.preprocess_fil_stage import PreprocessFILStage
 from morpheus.stages.preprocess.preprocess_nlp_stage import PreprocessNLPStage
 from morpheus.utils.file_utils import load_labels_file
-from morpheus.utils.logger import configure_logging
+from morpheus.utils.logger import set_log_level
 
 E2E_CONFIG_FILE = os.path.join(TEST_DIRS.morpheus_root, "tests/benchmarks/e2e_test_configs.json")
 with open(E2E_CONFIG_FILE, 'r', encoding='UTF-8') as f:
@@ -44,7 +44,7 @@
 
 def nlp_pipeline(config: Config, input_file, repeat, vocab_hash_file, output_file, model_name):
 
-    configure_logging(log_level=logging.INFO)
+    set_log_level(log_level=logging.DEBUG)
 
     pipeline = LinearPipeline(config)
     pipeline.set_source(FileSourceStage(config, filename=input_file, repeat=repeat))
@@ -71,7 +71,7 @@ def nlp_pipeline(config: Config, input_file, repeat, vocab_hash_file, output_fil
 
 def fil_pipeline(config: Config, input_file, repeat, output_file, model_name):
 
-    configure_logging(log_level=logging.INFO)
+    set_log_level(log_level=logging.DEBUG)
 
     pipeline = LinearPipeline(config)
     pipeline.set_source(FileSourceStage(config, filename=input_file, repeat=repeat))

tests/benchmarks/test_bench_rag_standalone_pipeline.py

@@ -135,10 +135,12 @@ def test_rag_standalone_pipe(benchmark: collections.abc.Callable[[collections.ab
                              repeat_count: int,
                              import_mod: types.ModuleType,
                              llm_service_name: str):
+    if llm_service_name == "openai":
+        os.environ.update({"OPENAI_API_KEY": "test_api_key"})
     collection_name = f"test_bench_rag_standalone_pipe_{llm_service_name}"
     populate_milvus(milvus_server_uri=milvus_server_uri,
                     collection_name=collection_name,
-                    resource_kwargs=import_mod.build_milvus_config(embedding_size=EMBEDDING_SIZE),
+                    resource_kwargs=import_mod.build_default_milvus_config(embedding_size=EMBEDDING_SIZE),
                     df=dataset["service/milvus_rss_data.json"],
                     overwrite=True)
 

tests/benchmarks/test_bench_vdb_upload_pipeline.py

@@ -79,7 +79,7 @@ def _run_pipeline(config: Config,
     pipe.add_stage(
         WriteToVectorDBStage(config,
                              resource_name=collection_name,
-                             resource_kwargs=utils_mod.build_milvus_config(embedding_size=EMBEDDING_SIZE),
+                             resource_kwargs=utils_mod.build_default_milvus_config(embedding_size=EMBEDDING_SIZE),
                              recreate=True,
                              service="milvus",
                              uri=milvus_server_uri))
@@ -92,7 +92,7 @@ def _run_pipeline(config: Config,
 @pytest.mark.benchmark
 @pytest.mark.import_mod([
     os.path.join(TEST_DIRS.examples_dir, 'llm/common/utils.py'),
-    os.path.join(TEST_DIRS.examples_dir, 'llm/common/web_scraper_stage.py'),
+    os.path.join(TEST_DIRS.examples_dir, 'llm/vdb_upload/module/web_scraper_stage.py'),
 ])
 @mock.patch('feedparser.http.get')
 @mock.patch('requests.Session')