dsr1-fp4-1k1k-mid-curve

concurrency 8,192 · 20 hours ago

cron

passed

re-run

Commit

SHA6cd8d28

Vectorize _create_custom_4d_mask in CustomQwen2Decoder (#27527)

ckvermaAI·22 hours ago

PR #27527 · Vectorize _create_custom_4d_mask in CustomQwen2Decoder

Run

GitHub Actions27999739491-1

Slurm job5335

GPUs48·prefill 16 / decode 32

ISL / OSL1024 / 1024

Logscron/27999739491-1/1k1k/dsr1-fp4-1k1k-mid-curve/2026-06-23/5335/sa-bench_isl_1024_osl_1024/

Metrics

28 captured

best_of

1.00

burstiness

1.00

completed

81,920

duration

727

max_concurrency

8,192

mean_e2el_ms

70,249ms

mean_itl_ms

975ms

mean_tpot_ms

19.91ms

mean_ttft_ms

51,922ms

median_e2el_ms

73,649ms

median_itl_ms

803ms

median_tpot_ms

17.33ms

median_ttft_ms

58,138ms

num_prompts

81,920

output_throughput

103,838tok/s

p99_e2el_ms

84,485ms

p99_itl_ms

3,142ms

p99_tpot_ms

39.04ms

p99_ttft_ms

64,405ms

peak_output_tokens_per_s

150,039s

request_throughput

113tok/s

std_e2el_ms

9,344ms

std_itl_ms

577ms

std_tpot_ms

7.41ms

std_ttft_ms

12,939ms

total_input_tokens

75,511,905

total_output_tokens

75,496,082

total_token_throughput

207,698tok/s