dsr1-fp8-1k1k-max-tpt

concurrency 2,048 · 11 days ago

cron

passed

re-run

Commit

SHAa0c6e0b

chore: bump tokenspeed_mla 0.1.1 -> 0.1.6 (#28116)

kpham-sgl·11 days ago

PR #28116 · chore: bump tokenspeed_mla 0.1.1 -> 0.1.6

Run

GitHub Actions27455007552-1

Slurm job5303

GPUs48·prefill 16 / decode 32

ISL / OSL1024 / 1024

Logscron/27455007552-1/1k1k/dsr1-fp8-1k1k-max-tpt/2026-06-13/5303/sa-bench_isl_1024_osl_1024/

Metrics

28 captured

best_of

1.00

burstiness

1.00

completed

20,480

duration

284

max_concurrency

2,048

mean_e2el_ms

27,267ms

mean_itl_ms

1,343ms

mean_tpot_ms

27.50ms

mean_ttft_ms

1,930ms

median_e2el_ms

26,831ms

median_itl_ms

1,380ms

median_tpot_ms

27.72ms

median_ttft_ms

989ms

num_prompts

20,480

output_throughput

66,409tok/s

p99_e2el_ms

37,897ms

p99_itl_ms

1,804ms

p99_tpot_ms

28.61ms

p99_ttft_ms

12,347ms

peak_output_tokens_per_s

80,857s

request_throughput

72.00tok/s

std_e2el_ms

2,961ms

std_itl_ms

235ms

std_tpot_ms

0.88ms

std_ttft_ms

2,668ms

total_input_tokens

18,880,692

total_output_tokens

18,888,974

total_token_throughput

132,789tok/s