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

minor: add random flashinfer vs triton use case (#2409)

Browse Source
This commit is contained in:
Yineng Zhang
2024-12-09 04:15:21 +08:00
committed by GitHub
parent 74bc9184c3
commit f62055b528
1 changed files with 25 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
test/srt/configs/random_flashinfer_vs_triton_config.yaml Normal file
View File

@@ -0,0 +1,25 @@
tasks:
- name: sglang-128-4
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 128 --random-output 4 --request-rate 24 --num-prompt 1440
- name: sglang-triton-128-4
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache --attention-backend triton
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 128 --random-output 4 --request-rate 24 --num-prompt 1440
- name: sglang-2000-100
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 2000 --random-output 100 --request-rate 2 --num-prompt 120
- name: sglang-triton-2000-100
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache --attention-backend triton
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 2000 --random-output 100 --request-rate 2 --num-prompt 120
- name: sglang-4000-200
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 4000 --random-output 200 --request-rate 8 --num-prompt 480
- name: sglang-triton-4000-200
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache --attention-backend triton
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 4000 --random-output 200 --request-rate 8 --num-prompt 480
- name: sglang-32000-100
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 32000 --random-output 100 --request-rate 1 --num-prompt 60
- name: sglang-triton-32000-100
server_cmd: python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --disable-radix-cache --attention-backend triton
client_cmd: python3 -m sglang.bench_serving --backend sglang --dataset-name random --random-input 32000 --random-output 100 --request-rate 1 --num-prompt 60