mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-10-04 04:04:14 +00:00
48a551ecbc
Some checks failed
Integration Tests (Replay) / Integration Tests (, , , client=, ) (push) Failing after 3s
Unit Tests / unit-tests (3.13) (push) Failing after 3s
Update ReadTheDocs / update-readthedocs (push) Failing after 3s
Test Llama Stack Build / build (push) Failing after 3s
Integration Auth Tests / test-matrix (oauth2_token) (push) Failing after 1s
SqlStore Integration Tests / test-postgres (3.12) (push) Failing after 1s
SqlStore Integration Tests / test-postgres (3.13) (push) Failing after 0s
Test External Providers Installed via Module / test-external-providers-from-module (venv) (push) Has been skipped
Test Llama Stack Build / generate-matrix (push) Successful in 3s
Python Package Build Test / build (3.12) (push) Failing after 1s
Python Package Build Test / build (3.13) (push) Failing after 2s
Test Llama Stack Build / build-custom-container-distribution (push) Failing after 3s
Test Llama Stack Build / build-single-provider (push) Failing after 3s
Vector IO Integration Tests / test-matrix (push) Failing after 5s
Test Llama Stack Build / build-ubi9-container-distribution (push) Failing after 4s
API Conformance Tests / check-schema-compatibility (push) Successful in 8s
Test External API and Providers / test-external (venv) (push) Failing after 3s
Unit Tests / unit-tests (3.12) (push) Failing after 4s
UI Tests / ui-tests (22) (push) Successful in 40s
Pre-commit / pre-commit (push) Successful in 1m9s
# What does this PR do? - Mostly AI-generated scripts to run guidellm (https://github.com/vllm-project/guidellm) benchmarks on k8s setup - Stack is using image built from main on 9/11 ## Test Plan See updated README.md
171 lines
13 KiB
Text
171 lines
13 KiB
Text
Collecting uv
|
|
Downloading uv-0.8.19-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (11 kB)
|
|
Downloading uv-0.8.19-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (20.9 MB)
|
|
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 20.9/20.9 MB 156.8 MB/s eta 0:00:00
|
|
Installing collected packages: uv
|
|
Successfully installed uv-0.8.19
|
|
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
|
|
|
|
[notice] A new release of pip is available: 24.0 -> 25.2
|
|
[notice] To update, run: pip install --upgrade pip
|
|
Using Python 3.11.13 environment at: /usr/local
|
|
Resolved 61 packages in 480ms
|
|
Downloading pillow (6.3MiB)
|
|
Downloading pydantic-core (1.9MiB)
|
|
Downloading pyarrow (40.8MiB)
|
|
Downloading aiohttp (1.7MiB)
|
|
Downloading numpy (16.2MiB)
|
|
Downloading pygments (1.2MiB)
|
|
Downloading transformers (11.1MiB)
|
|
Downloading pandas (11.8MiB)
|
|
Downloading tokenizers (3.1MiB)
|
|
Downloading hf-xet (3.0MiB)
|
|
Downloading pydantic-core
|
|
Downloading aiohttp
|
|
Downloading tokenizers
|
|
Downloading hf-xet
|
|
Downloading pygments
|
|
Downloading pillow
|
|
Downloading numpy
|
|
Downloading pandas
|
|
Downloading pyarrow
|
|
Downloading transformers
|
|
Prepared 61 packages in 1.25s
|
|
Installed 61 packages in 126ms
|
|
+ aiohappyeyeballs==2.6.1
|
|
+ aiohttp==3.12.15
|
|
+ aiosignal==1.4.0
|
|
+ annotated-types==0.7.0
|
|
+ anyio==4.10.0
|
|
+ attrs==25.3.0
|
|
+ certifi==2025.8.3
|
|
+ charset-normalizer==3.4.3
|
|
+ click==8.1.8
|
|
+ datasets==4.1.1
|
|
+ dill==0.4.0
|
|
+ filelock==3.19.1
|
|
+ frozenlist==1.7.0
|
|
+ fsspec==2025.9.0
|
|
+ ftfy==6.3.1
|
|
+ guidellm==0.3.0
|
|
+ h11==0.16.0
|
|
+ h2==4.3.0
|
|
+ hf-xet==1.1.10
|
|
+ hpack==4.1.0
|
|
+ httpcore==1.0.9
|
|
+ httpx==0.28.1
|
|
+ huggingface-hub==0.35.0
|
|
+ hyperframe==6.1.0
|
|
+ idna==3.10
|
|
+ loguru==0.7.3
|
|
+ markdown-it-py==4.0.0
|
|
+ mdurl==0.1.2
|
|
+ multidict==6.6.4
|
|
+ multiprocess==0.70.16
|
|
+ numpy==2.3.3
|
|
+ packaging==25.0
|
|
+ pandas==2.3.2
|
|
+ pillow==11.3.0
|
|
+ propcache==0.3.2
|
|
+ protobuf==6.32.1
|
|
+ pyarrow==21.0.0
|
|
+ pydantic==2.11.9
|
|
+ pydantic-core==2.33.2
|
|
+ pydantic-settings==2.10.1
|
|
+ pygments==2.19.2
|
|
+ python-dateutil==2.9.0.post0
|
|
+ python-dotenv==1.1.1
|
|
+ pytz==2025.2
|
|
+ pyyaml==6.0.2
|
|
+ regex==2025.9.18
|
|
+ requests==2.32.5
|
|
+ rich==14.1.0
|
|
+ safetensors==0.6.2
|
|
+ six==1.17.0
|
|
+ sniffio==1.3.1
|
|
+ tokenizers==0.22.1
|
|
+ tqdm==4.67.1
|
|
+ transformers==4.56.2
|
|
+ typing-extensions==4.15.0
|
|
+ typing-inspection==0.4.1
|
|
+ tzdata==2025.2
|
|
+ urllib3==2.5.0
|
|
+ wcwidth==0.2.14
|
|
+ xxhash==3.5.0
|
|
+ yarl==1.20.1
|
|
Using Python 3.11.13 environment at: /usr/local
|
|
Audited 1 package in 4ms
|
|
Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
|
|
Creating backend...
|
|
Backend openai_http connected to http://llama-stack-benchmark-service:8323/v1/openai for model meta-llama/Llama-3.2-3B-Instruct.
|
|
Creating request loader...
|
|
Created loader with 1000 unique requests from prompt_tokens=512,output_tokens=256.
|
|
|
|
|
|
╭─ Benchmarks ─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
|
|
│ [17:55:59] ⠋ 100% concurrent@1 (complete) Req: 0.3 req/s, 3.33s Lat, 1.0 Conc, 18 Comp, 1 Inc, 0 Err │
|
|
│ Tok: 74.0 gen/s, 238.0 tot/s, 49.6ms TTFT, 13.4ms ITL, 546 Prompt, 246 Gen │
|
|
│ [17:57:04] ⠋ 100% concurrent@2 (complete) Req: 0.6 req/s, 3.32s Lat, 1.9 Conc, 35 Comp, 2 Inc, 0 Err │
|
|
│ Tok: 137.1 gen/s, 457.5 tot/s, 50.6ms TTFT, 14.0ms ITL, 546 Prompt, 234 Gen │
|
|
│ [17:58:09] ⠋ 100% concurrent@4 (complete) Req: 1.2 req/s, 3.42s Lat, 4.0 Conc, 69 Comp, 4 Inc, 0 Err │
|
|
│ Tok: 276.7 gen/s, 907.2 tot/s, 52.7ms TTFT, 14.1ms ITL, 547 Prompt, 240 Gen │
|
|
│ [17:59:14] ⠋ 100% concurrent@8 (complete) Req: 2.3 req/s, 3.47s Lat, 7.8 Conc, 134 Comp, 8 Inc, 0 Err │
|
|
│ Tok: 541.4 gen/s, 1775.4 tot/s, 57.3ms TTFT, 14.3ms ITL, 547 Prompt, 240 Gen │
|
|
│ [18:00:19] ⠋ 100% concurrent@16 (complete) Req: 4.3 req/s, 3.60s Lat, 15.6 Conc, 259 Comp, 16 Inc, 0 Err │
|
|
│ Tok: 1034.8 gen/s, 3401.7 tot/s, 72.3ms TTFT, 14.8ms ITL, 547 Prompt, 239 Gen │
|
|
│ [18:01:25] ⠋ 100% concurrent@32 (complete) Req: 8.4 req/s, 3.69s Lat, 31.1 Conc, 505 Comp, 32 Inc, 0 Err │
|
|
│ Tok: 2029.7 gen/s, 6641.5 tot/s, 91.6ms TTFT, 15.0ms ITL, 547 Prompt, 241 Gen │
|
|
│ [18:02:31] ⠋ 100% concurrent@64 (complete) Req: 13.6 req/s, 4.50s Lat, 61.4 Conc, 818 Comp, 64 Inc, 0 Err │
|
|
│ Tok: 3333.9 gen/s, 10787.0 tot/s, 171.3ms TTFT, 17.8ms ITL, 547 Prompt, 244 Gen │
|
|
│ [18:03:40] ⠋ 100% concurrent@128 (complete) Req: 16.1 req/s, 7.43s Lat, 119.5 Conc, 964 Comp, 122 Inc, 0 Err │
|
|
│ Tok: 3897.0 gen/s, 12679.4 tot/s, 446.4ms TTFT, 28.9ms ITL, 547 Prompt, 243 Gen │
|
|
╰──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
|
|
Generating... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ (8/8) [ 0:08:41 < 0:00:00 ]
|
|
|
|
Benchmarks Metadata:
|
|
Run id:5393e64f-d9f8-4548-95d8-da320bba1c24
|
|
Duration:530.1 seconds
|
|
Profile:type=concurrent, strategies=['concurrent', 'concurrent', 'concurrent', 'concurrent', 'concurrent', 'concurrent', 'concurrent', 'concurrent'], streams=[1, 2, 4, 8, 16, 32, 64, 128]
|
|
Args:max_number=None, max_duration=60.0, warmup_number=None, warmup_duration=3.0, cooldown_number=None, cooldown_duration=None
|
|
Worker:type_='generative_requests_worker' backend_type='openai_http' backend_target='http://llama-stack-benchmark-service:8323/v1/openai' backend_model='meta-llama/Llama-3.2-3B-Instruct'
|
|
backend_info={'max_output_tokens': 16384, 'timeout': 300, 'http2': True, 'follow_redirects': True, 'headers': {}, 'text_completions_path': '/v1/completions', 'chat_completions_path':
|
|
'/v1/chat/completions'}
|
|
Request Loader:type_='generative_request_loader' data='prompt_tokens=512,output_tokens=256' data_args=None processor='meta-llama/Llama-3.2-3B-Instruct' processor_args=None
|
|
Extras:None
|
|
|
|
|
|
Benchmarks Info:
|
|
===================================================================================================================================================
|
|
Metadata |||| Requests Made ||| Prompt Tok/Req ||| Output Tok/Req ||| Prompt Tok Total||| Output Tok Total||
|
|
Benchmark| Start Time| End Time| Duration (s)| Comp| Inc| Err| Comp| Inc| Err| Comp| Inc| Err| Comp| Inc| Err| Comp| Inc| Err
|
|
--------------|-----------|---------|-------------|------|-----|-----|------|------|----|------|------|----|-------|------|----|-------|------|----
|
|
concurrent@1| 17:56:04| 17:57:04| 60.0| 18| 1| 0| 546.4| 512.0| 0.0| 246.4| 256.0| 0.0| 9836| 512| 0| 4436| 256| 0
|
|
concurrent@2| 17:57:09| 17:58:09| 60.0| 35| 2| 0| 546.4| 512.0| 0.0| 233.9| 132.0| 0.0| 19124| 1024| 0| 8188| 264| 0
|
|
concurrent@4| 17:58:14| 17:59:14| 60.0| 69| 4| 0| 546.6| 512.0| 0.0| 239.9| 60.5| 0.0| 37715| 2048| 0| 16553| 242| 0
|
|
concurrent@8| 17:59:19| 18:00:19| 60.0| 134| 8| 0| 546.6| 512.0| 0.0| 239.8| 126.6| 0.0| 73243| 4096| 0| 32135| 1013| 0
|
|
concurrent@16| 18:00:24| 18:01:24| 60.0| 259| 16| 0| 546.6| 512.0| 0.0| 239.0| 115.7| 0.0| 141561| 8192| 0| 61889| 1851| 0
|
|
concurrent@32| 18:01:30| 18:02:30| 60.0| 505| 32| 0| 546.5| 512.0| 0.0| 240.5| 113.2| 0.0| 275988| 16384| 0| 121466| 3623| 0
|
|
concurrent@64| 18:02:37| 18:03:37| 60.0| 818| 64| 0| 546.6| 512.0| 0.0| 244.5| 132.4| 0.0| 447087| 32768| 0| 199988| 8475| 0
|
|
concurrent@128| 18:03:45| 18:04:45| 60.0| 964| 122| 0| 546.5| 512.0| 0.0| 242.5| 133.1| 0.0| 526866| 62464| 0| 233789| 16241| 0
|
|
===================================================================================================================================================
|
|
|
|
|
|
Benchmarks Stats:
|
|
=======================================================================================================================================================
|
|
Metadata | Request Stats || Out Tok/sec| Tot Tok/sec| Req Latency (sec) ||| TTFT (ms) ||| ITL (ms) ||| TPOT (ms) ||
|
|
Benchmark| Per Second| Concurrency| mean| mean| mean| median| p99| mean| median| p99| mean| median| p99| mean| median| p99
|
|
--------------|-----------|------------|------------|------------|------|--------|------|------|-------|-------|-----|-------|-----|-----|-------|-----
|
|
concurrent@1| 0.30| 1.00| 74.0| 238.0| 3.33| 3.44| 3.63| 49.6| 47.2| 66.1| 13.4| 13.3| 14.0| 13.3| 13.3| 14.0
|
|
concurrent@2| 0.59| 1.95| 137.1| 457.5| 3.32| 3.61| 3.67| 50.6| 48.6| 80.4| 14.0| 14.0| 14.2| 13.9| 13.9| 14.1
|
|
concurrent@4| 1.15| 3.95| 276.7| 907.2| 3.42| 3.61| 3.77| 52.7| 49.7| 106.9| 14.1| 14.0| 14.6| 14.0| 13.9| 14.5
|
|
concurrent@8| 2.26| 7.83| 541.4| 1775.4| 3.47| 3.70| 3.79| 57.3| 50.9| 171.3| 14.3| 14.3| 14.4| 14.2| 14.2| 14.4
|
|
concurrent@16| 4.33| 15.57| 1034.8| 3401.7| 3.60| 3.81| 4.22| 72.3| 52.0| 292.9| 14.8| 14.7| 16.3| 14.7| 14.7| 16.3
|
|
concurrent@32| 8.44| 31.12| 2029.7| 6641.5| 3.69| 3.89| 4.24| 91.6| 62.6| 504.6| 15.0| 15.0| 15.4| 14.9| 14.9| 15.4
|
|
concurrent@64| 13.64| 61.40| 3333.9| 10787.0| 4.50| 4.61| 5.67| 171.3| 101.2| 1165.6| 17.8| 17.7| 19.2| 17.7| 17.6| 19.1
|
|
concurrent@128| 16.07| 119.45| 3897.0| 12679.4| 7.43| 7.63| 9.74| 446.4| 195.8| 2533.1| 28.9| 28.9| 31.0| 28.8| 28.8| 30.9
|
|
=======================================================================================================================================================
|
|
|
|
Saving benchmarks report...
|
|
Benchmarks report saved to /benchmarks.json
|
|
|
|
Benchmarking complete.
|