{ "label": "", "model": "sensenova-loop-bf16", "base_url": "http://127.0.0.1:30080/v1", "host": "tux", "time": "2026-10-30T16:33:57-0700", "config": { "base_url": "http://127.0.0.1:30080/v1", "model": "sensenova-loop-bf16", "concurrency": "1,2,4,8", "num_requests": 8, "warmup": 2, "input_tokens": 1024, "output_tokens": 256, "tokenizer": "upstream/SenseNova-Loop-9B-Preview/tokenizer.json", "allow_prefix_cache": false, "no_ignore_eos": false, "timeout": 600, "label": "", "out": "eval/runs/perf_sglang_bf16_mtp.json", "prompt_text_tokens": 1024, "prompt_build": "tokenizer.json" }, "levels": [ { "concurrency": 1, "n_requests": 8, "n_ok": 8, "n_errors": 0, "errors": [], "wall_s": 272.73132864851505, "ttft_s": { "p50": 0.7848195834085345, "p90": 0.9759742130525411, "mean": 0.8312031434616074 }, "e2e_s": { "p50": 34.21052242349833, "p90": 34.438923329673706, "mean": 34.09125282650348 }, "decode_tok_s_per_req": { "p50": 7.627452517131835, "p90": 7.763279273008121, "mean": 7.669396340870895 }, "itl_s": { "p50": 0.1311053762084567, "p90": 0.13172920723317885, "mean": 0.13043050938103276 }, "output_tokens_mean": 256.0, "output_tokens_min": 256, "output_tokens_max": 256, "prompt_tokens_mean": 1051.0, "n_short_outputs": 0, "aggregate_output_tok_s": 7.50922165835733, "requests_per_s": 0.02933289710295832, "usage_reported": true }, { "concurrency": 2, "n_requests": 8, "n_ok": 8, "n_errors": 0, "errors": [], "wall_s": 138.61350492294878, "ttft_s": { "p50": 1.1976498886942863, "p90": 1.6063555277884007, "mean": 1.2973694901447743 }, "e2e_s": { "p50": 34.72376540116966, "p90": 36.45519359242171, "mean": 34.5663175820373 }, "decode_tok_s_per_req": { "p50": 7.653260406224514, "p90": 8.09628260642012, "mean": 7.677757556692128 }, "itl_s": { "p50": 0.13067035353651235, "p90": 0.13736926173243452, "mean": 0.13046496708274763 }, "output_tokens_mean": 256.0, "output_tokens_min": 256, "output_tokens_max": 256, "prompt_tokens_mean": 1051.0, "n_short_outputs": 0, "aggregate_output_tok_s": 14.774895138380808, "requests_per_s": 0.05771443413430003, "usage_reported": true }, { "concurrency": 4, "n_requests": 8, "n_ok": 8, "n_errors": 0, "errors": [], "wall_s": 76.38755708001554, "ttft_s": { "p50": 2.5858272635377944, "p90": 2.779972269944847, "mean": 2.2637667186791077 }, "e2e_s": { "p50": 36.801751386839896, "p90": 39.171385447308424, "mean": 37.270905990269966 }, "decode_tok_s_per_req": { "p50": 7.2633816269398785, "p90": 7.592177491148207, "mean": 7.294118130503968 }, "itl_s": { "p50": 0.1376772181057901, "p90": 0.1427121850625411, "mean": 0.13728188767883123 }, "output_tokens_mean": 256.0, "output_tokens_min": 256, "output_tokens_max": 256, "prompt_tokens_mean": 1051.0, "n_short_outputs": 0, "aggregate_output_tok_s": 26.810649250829314, "requests_per_s": 0.10472909863605201, "usage_reported": true }, { "concurrency": 8, "n_requests": 8, "n_ok": 8, "n_errors": 0, "errors": [], "wall_s": 44.81087386980653, "ttft_s": { "p50": 5.374245874583721, "p90": 5.41043642014265, "mean": 5.388463915442117 }, "e2e_s": { "p50": 41.73927395185456, "p90": 43.86301381308586, "mean": 41.677275480818935 }, "decode_tok_s_per_req": { "p50": 7.0119672437329355, "p90": 7.471562651588895, "mean": 7.0494192349641605 }, "itl_s": { "p50": 0.1426148278529153, "p90": 0.1509358527029262, "mean": 0.14230802183349928 }, "output_tokens_mean": 256.0, "output_tokens_min": 256, "output_tokens_max": 256, "prompt_tokens_mean": 1051.0, "n_short_outputs": 0, "aggregate_output_tok_s": 45.7031926212878, "requests_per_s": 0.17852809617690546, "usage_reported": true } ] }