sglang/ GB200 perf

dsr1-fp4-1k1k-mid-curve

concurrency 4,096 · 21 days ago

cron
passed
re-run
Commit

[diffusion] fix: fix lingbot realtime consistency gt pin (#27080)

mickqian·21 days ago
PR #27080 · [diffusion] Fix LingBot realtime consistency GT pin
Run
GitHub Actions26861695033-1
Slurm job5112
GPUs48·prefill 16 / decode 32
ISL / OSL1024 / 1024

Metrics

28 captured

best_of

1.00

burstiness

1.00

completed

40,960

duration

305

max_concurrency

4,096

mean_e2el_ms

29,833ms

mean_itl_ms

1,103ms

mean_tpot_ms

22.51ms

mean_ttft_ms

9,113ms

median_e2el_ms

29,777ms

median_itl_ms

1,098ms

median_tpot_ms

22.94ms

median_ttft_ms

9,507ms

num_prompts

40,960

output_throughput

123,548tok/s

p99_e2el_ms

47,170ms

p99_itl_ms

2,263ms

p99_tpot_ms

27.51ms

p99_ttft_ms

23,640ms

peak_output_tokens_per_s

181,243s

request_throughput

134tok/s

std_e2el_ms

4,314ms

std_itl_ms

342ms

std_tpot_ms

3.35ms

std_ttft_ms

4,024ms

total_input_tokens

37,769,666

total_output_tokens

37,742,239

total_token_throughput

247,185tok/s