{ "label": "", "model": "sensenova-loop-nvfp4", "base_url": "http://127.0.0.1:30080/v1", "host": "tux", "time": "2026-10-30T17:00:57-0700", "config": { "base_url": "http://127.0.0.1:30080/v1", "model": "sensenova-loop-nvfp4", "concurrency": "1,2,4,8", "num_requests": 16, "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_nvfp4_v1.json", "prompt_text_tokens": 1024, "prompt_build": "tokenizer.json" }, "levels": [ { "concurrency": 1, "n_requests": 16, "n_ok": 16, "n_errors": 0, "errors": [], "wall_s": 436.37869714573026, "ttft_s": { "p50": 0.5826272442936897, "p90": 0.5880275997333229, "mean": 0.5825185888679698 }, "e2e_s": { "p50": 27.272629408165812, "p90": 27.33341192267835, "mean": 27.273564630304463 }, "decode_tok_s_per_req": { "p50": 9.555136327634862, "p90": 9.571729349173973, "mean": 9.55383492852507 }, "itl_s": { "p50": 0.10465575453697466, "p90": 0.10489086669814937, "mean": 0.10467025827383622 }, "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": 9.386342703690978, "requests_per_s": 0.036665401186292884, "usage_reported": true }, { "concurrency": 2, "n_requests": 16, "n_ok": 16, "n_errors": 0, "errors": [], "wall_s": 232.5841392017901, "ttft_s": { "p50": 1.1243764483369887, "p90": 1.1582226920872927, "mean": 1.117659447598271 }, "e2e_s": { "p50": 29.07659900886938, "p90": 29.158572073560208, "mean": 29.072850470547564 }, "decode_tok_s_per_req": { "p50": 9.119639853361988, "p90": 9.150102465833314, "mean": 9.121841848607799 }, "itl_s": { "p50": 0.1096534530486108, "p90": 0.10988617092918823, "mean": 0.10962762238854981 }, "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": 17.610831134303226, "requests_per_s": 0.06879230911837197, "usage_reported": true }, { "concurrency": 4, "n_requests": 16, "n_ok": 16, "n_errors": 0, "errors": [], "wall_s": 122.35653864778578, "ttft_s": { "p50": 2.017746684141457, "p90": 2.0371825736947358, "mean": 2.0016646458534524 }, "e2e_s": { "p50": 30.58990833349526, "p90": 30.627535844687372, "mean": 30.58872388175223 }, "decode_tok_s_per_req": { "p50": 8.92044357838708, "p90": 8.939535982689941, "mean": 8.920202608225608 }, "itl_s": { "p50": 0.11210204872540107, "p90": 0.11229545274091994, "mean": 0.1121052856137063 }, "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": 33.47593880365235, "requests_per_s": 0.13076538595176698, "usage_reported": true }, { "concurrency": 8, "n_requests": 16, "n_ok": 16, "n_errors": 0, "errors": [], "wall_s": 63.220375599339604, "ttft_s": { "p50": 4.205650780815631, "p90": 4.5575914164073765, "mean": 4.21742238936713 }, "e2e_s": { "p50": 31.60961770033464, "p90": 32.07297795498744, "mean": 31.608251873170957 }, "decode_tok_s_per_req": { "p50": 9.306845748863292, "p90": 9.356793657252013, "mean": 9.310041888446369 }, "itl_s": { "p50": 0.10744962619946283, "p90": 0.10790295124090478, "mean": 0.10741330675379027 }, "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": 64.78923861443788, "requests_per_s": 0.25308296333764796, "usage_reported": true } ] }