{ "profile": "gb10", "slo": { "ttft_ms": 1000.0, "tpot_ms": 100.0 }, "environment": { "sglang": "0.5.22", "sglang_kernel": "0.4.9", "flashinfer": "0.7.1", "torch": "2.14.1+cu130", "cuda": "13.0", "driver": "580.159.03", "gpu": "NVIDIA GB10", "compute_capability": "12.1", "gpu_memory_total_gib": 121.7, "gpu_memory_free_gib_at_start": 67.7, "python": "3.12.3", "machine": "aarch64", "profile": "gb10", "models": [ "Qwen/Qwen3-30B-A3B-Instruct-2507", "Qwen/Qwen3-30B-A3B-Instruct-2507-FP8", "lmsys/SGLang-EAGLE3-Qwen3-30B-A3B-Instruct-2507-SpecForge-Nex" ] }, "load": { "default": { "kv_pool_tokens": 234600, "points": [ { "x": 1.0, "output_throughput": { "median": 29.14, "min": 28.48, "max": 29.27, "n": 3 }, "median_ttft_ms": { "median": 272.26, "min": 269.57, "max": 279.58, "n": 3 }, "p99_ttft_ms": { "median": 339.83, "min": 308.77, "max": 359.69, "n": 3 }, "median_tpot_ms": { "median": 32.52, "min": 32.07, "max": 33.23, "n": 3 }, "p99_tpot_ms": { "median": 34.08, "min": 33.62, "max": 34.55, "n": 3 }, "goodput_rps": { "median": 0.23, "min": 0.22, "max": 0.23, "n": 3 }, "request_throughput": { "median": 0.23, "min": 0.22, "max": 0.23, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 42.6, 42.6, 43.3, 63.6, 92.4, 208.4, 210.3, 245.8, 248.6, 249.9, 251.4, 257.6, 258.6, 259.3, 261.4, 263.0, 264.0, 264.0, 267.7, 268.7, 269.6, 270.4, 270.5, 270.9, 271.2, 272.3, 272.5, 274.0, 274.9, 276.0, 277.5, 278.6, 278.7, 279.6, 279.6, 280.3, 280.6, 282.5, 284.3, 285.6, 287.4, 288.3, 290.0, 291.5, 292.5, 292.5, 302.1, 302.9, 309.9, 346.4, 371.7 ], "tpot_quantiles_ms": [ 31.8, 31.8, 31.9, 31.9, 31.9, 32.0, 32.0, 32.0, 32.0, 32.1, 32.1, 32.1, 32.2, 32.2, 32.2, 32.2, 32.3, 32.3, 32.4, 32.4, 32.4, 32.5, 32.5, 32.5, 32.6, 32.6, 32.6, 32.7, 32.7, 32.7, 32.8, 32.9, 33.1, 33.1, 33.1, 33.1, 33.2, 33.2, 33.3, 33.3, 33.3, 33.4, 33.4, 33.5, 33.5, 33.6, 33.6, 33.7, 34.1, 34.2, 34.6 ] }, { "x": 2.0, "output_throughput": { "median": 43.83, "min": 43.65, "max": 44.86, "n": 3 }, "median_ttft_ms": { "median": 69.19, "min": 68.56, "max": 71.33, "n": 3 }, "p99_ttft_ms": { "median": 78.22, "min": 76.68, "max": 81.78, "n": 3 }, "median_tpot_ms": { "median": 45.42, "min": 44.73, "max": 45.49, "n": 3 }, "p99_tpot_ms": { "median": 47.28, "min": 47.22, "max": 47.78, "n": 3 }, "goodput_rps": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "request_throughput": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 61.3, 61.3, 61.3, 61.9, 62.6, 63.0, 63.0, 63.2, 63.2, 63.4, 63.5, 63.7, 63.8, 64.0, 64.0, 64.1, 64.2, 64.2, 64.4, 64.8, 64.9, 64.9, 65.0, 65.7, 66.0, 66.9, 73.0, 73.4, 73.5, 73.7, 74.0, 74.3, 74.5, 74.6, 74.6, 74.6, 74.8, 75.2, 75.2, 75.5, 75.8, 76.0, 76.0, 76.1, 76.4, 76.5, 76.8, 77.0, 77.9, 78.3, 82.9 ], "tpot_quantiles_ms": [ 41.2, 41.2, 41.3, 41.7, 41.9, 42.9, 43.0, 44.0, 44.1, 44.1, 44.2, 44.4, 44.5, 44.5, 44.5, 44.6, 44.7, 44.7, 44.7, 44.8, 44.9, 45.0, 45.0, 45.0, 45.0, 45.1, 45.1, 45.1, 45.2, 45.3, 45.5, 45.6, 45.7, 45.8, 45.8, 45.9, 45.9, 46.1, 46.1, 46.2, 46.2, 46.3, 46.4, 47.1, 47.2, 47.2, 47.2, 47.3, 47.3, 47.7, 47.8 ] }, { "x": 4.0, "output_throughput": { "median": 56.89, "min": 56.58, "max": 57.47, "n": 3 }, "median_ttft_ms": { "median": 113.49, "min": 112.83, "max": 114.97, "n": 3 }, "p99_ttft_ms": { "median": 115.79, "min": 115.7, "max": 117.2, "n": 3 }, "median_tpot_ms": { "median": 71.01, "min": 70.18, "max": 71.13, "n": 3 }, "p99_tpot_ms": { "median": 73.16, "min": 72.0, "max": 74.07, "n": 3 }, "goodput_rps": { "median": 0.44, "min": 0.44, "max": 0.45, "n": 3 }, "request_throughput": { "median": 0.44, "min": 0.44, "max": 0.45, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 47.0, 47.0, 47.8, 48.4, 48.9, 49.5, 49.5, 49.9, 50.3, 50.4, 50.6, 50.7, 51.6, 110.3, 110.6, 110.9, 112.7, 112.7, 113.0, 113.3, 113.3, 113.4, 113.4, 113.4, 113.5, 113.5, 113.5, 113.5, 113.6, 114.5, 114.6, 114.7, 114.8, 114.8, 114.8, 114.8, 114.9, 115.0, 115.2, 115.4, 115.5, 115.6, 115.6, 115.7, 115.8, 116.1, 116.2, 116.2, 117.0, 117.1, 117.2 ], "tpot_quantiles_ms": [ 64.3, 64.3, 64.3, 64.3, 64.8, 65.5, 65.5, 65.5, 66.0, 67.1, 67.1, 67.1, 67.6, 69.0, 69.0, 69.0, 69.5, 69.5, 69.7, 69.7, 69.7, 69.7, 69.7, 69.7, 70.2, 70.2, 70.2, 70.2, 70.2, 70.7, 71.6, 71.6, 71.6, 72.1, 72.1, 72.1, 72.1, 72.1, 72.6, 72.6, 72.6, 72.6, 72.7, 72.7, 72.7, 73.0, 73.2, 73.6, 73.6, 73.6, 74.1 ] }, { "x": 8.0, "output_throughput": { "median": 74.11, "min": 70.55, "max": 77.57, "n": 3 }, "median_ttft_ms": { "median": 209.98, "min": 185.25, "max": 216.02, "n": 3 }, "p99_ttft_ms": { "median": 1164.8, "min": 1154.97, "max": 1256.42, "n": 3 }, "median_tpot_ms": { "median": 102.21, "min": 102.1, "max": 111.2, "n": 3 }, "p99_tpot_ms": { "median": 113.31, "min": 109.35, "max": 117.85, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.04, "n": 3 }, "request_throughput": { "median": 0.58, "min": 0.55, "max": 0.61, "n": 3 }, "completed": [ 32, 32, 32 ], "ttft_quantiles_ms": [ 47.5, 49.6, 51.2, 52.9, 148.1, 148.2, 148.6, 148.8, 148.9, 149.0, 149.1, 150.2, 151.0, 151.6, 152.2, 152.8, 152.8, 152.9, 159.9, 160.0, 160.1, 160.1, 167.0, 167.1, 167.2, 167.2, 264.8, 267.7, 302.8, 541.9, 596.6, 596.6, 620.3, 658.4, 941.2, 941.3, 941.3, 1085.1, 1085.6, 1086.2, 1152.0, 1152.1, 1152.3, 1154.0, 1154.6, 1155.0, 1164.1, 1164.8, 1255.5, 1255.9, 1256.5 ], "tpot_quantiles_ms": [ 84.7, 84.7, 84.7, 84.7, 91.7, 99.8, 99.8, 100.4, 100.4, 100.4, 100.4, 101.2, 101.3, 101.3, 101.3, 102.2, 102.2, 102.2, 102.2, 103.0, 103.0, 103.0, 104.0, 106.3, 106.3, 106.3, 106.7, 106.7, 106.8, 106.8, 107.1, 107.1, 107.1, 107.1, 107.4, 107.5, 110.0, 110.6, 111.2, 111.2, 111.9, 111.9, 111.9, 111.9, 112.7, 112.7, 112.7, 112.7, 113.2, 115.9, 118.7 ] }, { "x": 16.0, "output_throughput": { "median": 101.09, "min": 99.5, "max": 103.9, "n": 3 }, "median_ttft_ms": { "median": 238.47, "min": 210.12, "max": 252.39, "n": 3 }, "p99_ttft_ms": { "median": 1721.73, "min": 1698.29, "max": 1743.34, "n": 3 }, "median_tpot_ms": { "median": 154.24, "min": 151.78, "max": 158.66, "n": 3 }, "p99_tpot_ms": { "median": 162.18, "min": 160.82, "max": 166.21, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.79, "min": 0.78, "max": 0.81, "n": 3 }, "completed": [ 64, 64, 64 ], "ttft_quantiles_ms": [ 49.2, 52.4, 186.3, 187.2, 188.5, 188.9, 189.0, 189.2, 189.3, 201.2, 201.5, 201.6, 201.9, 203.2, 206.3, 206.6, 206.6, 206.7, 209.9, 210.0, 210.1, 210.3, 210.7, 211.2, 211.8, 237.6, 277.7, 618.7, 667.1, 818.8, 818.9, 854.6, 873.6, 1639.7, 1639.8, 1640.0, 1668.6, 1668.9, 1677.1, 1677.5, 1678.4, 1679.1, 1696.1, 1697.1, 1698.1, 1721.5, 1721.6, 1721.8, 1742.4, 1743.1, 1743.4 ], "tpot_quantiles_ms": [ 136.7, 136.7, 136.7, 136.7, 137.8, 144.6, 144.6, 144.6, 144.6, 147.3, 147.3, 148.2, 148.2, 148.2, 151.8, 151.8, 151.8, 151.8, 151.8, 151.8, 152.1, 152.1, 152.1, 153.7, 154.2, 154.2, 154.3, 154.3, 154.4, 154.4, 154.4, 154.4, 154.6, 154.7, 157.1, 157.1, 157.1, 158.1, 158.5, 158.7, 158.7, 158.7, 158.7, 160.3, 160.3, 160.3, 160.3, 160.3, 160.6, 163.4, 168.2 ] }, { "x": 32.0, "output_throughput": { "median": 150.66, "min": 148.01, "max": 151.22, "n": 3 }, "median_ttft_ms": { "median": 416.37, "min": 338.06, "max": 434.64, "n": 3 }, "p99_ttft_ms": { "median": 3061.6, "min": 3018.5, "max": 3701.0, "n": 3 }, "median_tpot_ms": { "median": 201.91, "min": 201.43, "max": 206.9, "n": 3 }, "p99_tpot_ms": { "median": 224.81, "min": 216.96, "max": 228.96, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.18, "min": 1.16, "max": 1.18, "n": 3 }, "completed": [ 128, 128, 128 ], "ttft_quantiles_ms": [ 53.5, 187.7, 188.7, 204.6, 221.8, 222.2, 245.5, 246.0, 246.8, 247.0, 247.6, 248.6, 249.2, 257.7, 268.7, 268.8, 269.2, 284.8, 311.7, 338.0, 347.7, 347.8, 354.4, 409.8, 410.8, 434.6, 655.9, 707.4, 997.3, 1969.7, 1972.4, 1973.8, 1975.7, 1978.5, 2032.5, 2034.4, 2053.1, 2054.3, 2054.9, 2056.0, 2271.5, 2272.9, 2972.1, 3010.2, 3012.2, 3018.5, 3060.6, 3081.2, 3083.8, 3699.3, 3701.1 ], "tpot_quantiles_ms": [ 190.6, 190.6, 193.5, 193.5, 194.2, 194.2, 195.3, 195.3, 195.9, 196.4, 198.3, 198.8, 198.8, 199.7, 199.7, 200.7, 201.4, 201.9, 201.9, 201.9, 201.9, 202.4, 202.9, 202.9, 202.9, 203.4, 203.8, 203.8, 206.4, 206.5, 206.5, 206.9, 206.9, 207.1, 207.2, 207.2, 208.9, 211.1, 211.1, 212.1, 212.9, 212.9, 214.0, 214.8, 214.8, 214.8, 217.3, 218.7, 218.7, 224.8, 232.9 ] }, { "x": 64.0, "output_throughput": { "median": 235.38, "min": 232.77, "max": 241.17, "n": 3 }, "median_ttft_ms": { "median": 557.43, "min": 554.63, "max": 586.46, "n": 3 }, "p99_ttft_ms": { "median": 5769.85, "min": 5669.33, "max": 5796.86, "n": 3 }, "median_tpot_ms": { "median": 254.06, "min": 248.17, "max": 254.27, "n": 3 }, "p99_tpot_ms": { "median": 292.33, "min": 280.93, "max": 296.27, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.84, "min": 1.82, "max": 1.88, "n": 3 }, "completed": [ 256, 256, 256 ], "ttft_quantiles_ms": [ 62.0, 190.7, 198.7, 253.7, 272.8, 274.3, 284.6, 285.0, 316.0, 328.2, 399.7, 406.1, 425.4, 425.8, 502.1, 504.8, 507.4, 510.4, 513.9, 532.5, 532.7, 533.1, 554.6, 572.2, 583.8, 585.4, 718.4, 1898.4, 1932.6, 1954.5, 1976.8, 2073.2, 2189.4, 3222.8, 3237.1, 3243.9, 3301.3, 3303.3, 3403.7, 3516.0, 4529.5, 4563.6, 4567.3, 4652.7, 4743.8, 4885.0, 5529.2, 5668.2, 5765.7, 5787.2, 5796.9 ], "tpot_quantiles_ms": [ 238.5, 242.3, 244.3, 244.3, 244.3, 244.4, 244.4, 245.7, 246.5, 246.9, 246.9, 246.9, 247.2, 247.4, 247.4, 247.5, 248.2, 249.2, 249.3, 249.4, 249.4, 250.8, 252.3, 252.5, 252.5, 252.5, 253.1, 253.7, 254.0, 254.3, 255.1, 255.2, 255.2, 257.6, 260.0, 260.0, 260.5, 260.6, 264.5, 267.6, 267.8, 270.4, 270.9, 271.1, 274.9, 278.1, 281.3, 281.6, 285.5, 288.4, 301.7 ] } ] }, "eagle3": { "kv_pool_tokens": 222796, "points": [ { "x": 1.0, "output_throughput": { "median": 40.93, "min": 38.95, "max": 42.83, "n": 3 }, "median_ttft_ms": { "median": 320.5, "min": 301.76, "max": 329.4, "n": 3 }, "p99_ttft_ms": { "median": 360.04, "min": 354.8, "max": 360.27, "n": 3 }, "median_tpot_ms": { "median": 20.3, "min": 18.97, "max": 20.38, "n": 3 }, "p99_tpot_ms": { "median": 38.34, "min": 33.78, "max": 40.8, "n": 3 }, "goodput_rps": { "median": 0.32, "min": 0.3, "max": 0.33, "n": 3 }, "request_throughput": { "median": 0.32, "min": 0.3, "max": 0.33, "n": 3 }, "completed": [ 16, 16, 16 ], "accept_length": { "median": 2.73, "min": 2.71, "max": 2.76, "n": 3 }, "ttft_quantiles_ms": [ 71.4, 71.4, 74.1, 78.6, 143.5, 264.3, 265.0, 269.8, 272.0, 272.3, 277.1, 288.1, 289.5, 289.9, 292.7, 299.8, 300.1, 300.1, 302.3, 302.5, 303.5, 307.9, 310.8, 312.5, 317.1, 318.0, 319.4, 321.6, 324.2, 324.9, 325.9, 326.9, 326.9, 327.3, 327.3, 328.4, 331.9, 332.4, 332.6, 335.2, 336.2, 338.0, 338.0, 339.8, 344.6, 347.5, 347.8, 354.5, 354.9, 362.5, 363.6 ], "tpot_quantiles_ms": [ 10.5, 10.5, 11.0, 14.6, 14.8, 15.1, 15.9, 15.9, 16.1, 16.2, 16.3, 16.6, 16.7, 16.8, 16.9, 17.0, 17.2, 17.2, 17.3, 17.6, 17.9, 18.7, 18.9, 19.2, 19.3, 19.9, 20.1, 20.7, 20.7, 20.8, 21.1, 22.1, 22.6, 23.2, 23.2, 25.0, 25.1, 26.0, 26.4, 26.6, 27.0, 27.1, 30.2, 30.8, 32.0, 34.3, 36.0, 36.4, 38.7, 40.5, 40.9 ] }, { "x": 2.0, "output_throughput": { "median": 56.38, "min": 56.35, "max": 58.84, "n": 3 }, "median_ttft_ms": { "median": 190.48, "min": 188.65, "max": 190.73, "n": 3 }, "p99_ttft_ms": { "median": 247.47, "min": 223.29, "max": 255.47, "n": 3 }, "median_tpot_ms": { "median": 29.16, "min": 27.55, "max": 29.46, "n": 3 }, "p99_tpot_ms": { "median": 57.82, "min": 48.12, "max": 65.54, "n": 3 }, "goodput_rps": { "median": 0.44, "min": 0.44, "max": 0.46, "n": 3 }, "request_throughput": { "median": 0.44, "min": 0.44, "max": 0.46, "n": 3 }, "completed": [ 16, 16, 16 ], "accept_length": { "median": 2.71, "min": 2.7, "max": 2.76, "n": 3 }, "ttft_quantiles_ms": [ 122.0, 122.0, 124.8, 124.9, 127.6, 131.0, 133.9, 145.2, 155.9, 168.0, 170.5, 171.4, 173.7, 175.7, 175.9, 176.1, 177.8, 177.8, 180.2, 181.2, 183.3, 185.2, 186.5, 187.3, 187.5, 190.0, 190.1, 191.4, 194.5, 196.5, 198.4, 200.6, 202.2, 204.8, 204.8, 205.5, 206.1, 206.5, 207.6, 209.7, 210.9, 211.5, 213.6, 214.3, 216.8, 217.0, 219.6, 224.4, 231.0, 252.9, 259.8 ], "tpot_quantiles_ms": [ 12.3, 12.3, 19.5, 20.4, 21.7, 22.6, 22.6, 22.7, 23.1, 24.0, 24.0, 24.0, 24.1, 24.1, 24.1, 24.7, 25.6, 25.6, 25.7, 25.7, 26.1, 26.4, 27.0, 28.5, 28.7, 28.7, 29.1, 29.7, 29.8, 31.3, 32.0, 32.5, 32.5, 34.4, 34.4, 34.8, 37.1, 38.5, 39.1, 39.6, 39.9, 39.9, 42.2, 46.7, 47.7, 48.4, 53.0, 53.9, 58.5, 61.7, 66.2 ] }, { "x": 4.0, "output_throughput": { "median": 73.36, "min": 72.93, "max": 78.18, "n": 3 }, "median_ttft_ms": { "median": 296.55, "min": 273.41, "max": 300.0, "n": 3 }, "p99_ttft_ms": { "median": 309.07, "min": 308.63, "max": 318.27, "n": 3 }, "median_tpot_ms": { "median": 43.38, "min": 39.57, "max": 43.5, "n": 3 }, "p99_tpot_ms": { "median": 81.25, "min": 70.4, "max": 93.86, "n": 3 }, "goodput_rps": { "median": 0.57, "min": 0.57, "max": 0.61, "n": 3 }, "request_throughput": { "median": 0.57, "min": 0.57, "max": 0.61, "n": 3 }, "completed": [ 16, 16, 16 ], "accept_length": { "median": 2.71, "min": 2.69, "max": 2.76, "n": 3 }, "ttft_quantiles_ms": [ 160.1, 160.1, 160.2, 166.6, 166.7, 166.7, 167.4, 167.6, 167.6, 171.2, 202.3, 235.6, 258.4, 263.8, 264.3, 267.6, 279.2, 279.2, 280.9, 282.7, 285.4, 285.8, 290.9, 292.7, 294.5, 295.3, 296.4, 296.7, 298.1, 298.6, 298.9, 299.4, 299.9, 300.1, 300.1, 302.2, 302.8, 303.3, 304.0, 304.7, 306.2, 306.4, 306.6, 308.4, 308.7, 308.9, 308.9, 309.1, 318.1, 318.2, 318.3 ], "tpot_quantiles_ms": [ 22.4, 22.4, 24.0, 26.6, 29.1, 29.2, 30.0, 31.3, 33.2, 33.6, 35.0, 35.3, 35.5, 36.4, 36.5, 36.8, 36.9, 36.9, 37.8, 38.7, 39.2, 39.5, 39.6, 40.3, 41.8, 42.4, 43.1, 43.7, 44.6, 45.0, 46.2, 47.6, 47.8, 48.4, 48.4, 51.9, 54.5, 55.1, 57.2, 57.7, 58.9, 61.1, 61.3, 63.3, 64.7, 71.4, 77.2, 79.7, 81.5, 92.9, 94.0 ] }, { "x": 8.0, "output_throughput": { "median": 97.06, "min": 86.93, "max": 107.23, "n": 3 }, "median_ttft_ms": { "median": 481.82, "min": 454.5, "max": 511.4, "n": 3 }, "p99_ttft_ms": { "median": 700.43, "min": 665.09, "max": 845.18, "n": 3 }, "median_tpot_ms": { "median": 59.44, "min": 58.4, "max": 66.86, "n": 3 }, "p99_tpot_ms": { "median": 134.61, "min": 123.34, "max": 181.58, "n": 3 }, "goodput_rps": { "median": 0.62, "min": 0.57, "max": 0.79, "n": 3 }, "request_throughput": { "median": 0.76, "min": 0.68, "max": 0.84, "n": 3 }, "completed": [ 32, 32, 32 ], "accept_length": { "median": 2.74, "min": 2.72, "max": 2.8, "n": 3 }, "ttft_quantiles_ms": [ 208.8, 221.5, 228.9, 229.0, 229.0, 229.2, 232.3, 232.4, 232.4, 233.0, 233.2, 233.3, 233.6, 280.4, 283.6, 398.5, 405.8, 408.5, 414.1, 416.8, 417.6, 418.4, 419.5, 427.4, 446.8, 450.7, 512.9, 550.8, 561.1, 572.0, 577.6, 577.6, 581.4, 596.2, 601.4, 605.2, 613.7, 619.5, 622.6, 629.7, 633.0, 634.0, 643.0, 646.1, 651.3, 671.3, 680.3, 685.3, 700.7, 833.5, 847.4 ], "tpot_quantiles_ms": [ 41.5, 44.8, 45.9, 46.6, 47.8, 48.7, 49.5, 51.0, 51.6, 51.8, 52.2, 53.2, 53.7, 54.4, 55.1, 55.7, 56.0, 56.4, 58.1, 58.4, 58.7, 58.9, 59.7, 60.4, 61.5, 62.5, 63.6, 64.5, 65.3, 66.0, 67.0, 67.5, 70.1, 70.6, 73.5, 74.6, 78.6, 82.1, 84.4, 88.7, 92.7, 92.9, 93.7, 94.8, 100.2, 108.6, 113.3, 129.5, 133.9, 143.0, 193.0 ] }, { "x": 16.0, "output_throughput": { "median": 145.49, "min": 140.81, "max": 153.72, "n": 3 }, "median_ttft_ms": { "median": 579.73, "min": 567.26, "max": 585.19, "n": 3 }, "p99_ttft_ms": { "median": 1145.17, "min": 1088.46, "max": 1172.65, "n": 3 }, "median_tpot_ms": { "median": 82.62, "min": 82.59, "max": 84.43, "n": 3 }, "p99_tpot_ms": { "median": 188.2, "min": 170.53, "max": 228.31, "n": 3 }, "goodput_rps": { "median": 0.67, "min": 0.64, "max": 0.83, "n": 3 }, "request_throughput": { "median": 1.14, "min": 1.1, "max": 1.2, "n": 3 }, "completed": [ 64, 64, 64 ], "accept_length": { "median": 2.73, "min": 2.71, "max": 2.83, "n": 3 }, "ttft_quantiles_ms": [ 152.6, 268.3, 273.3, 273.4, 273.4, 278.0, 278.1, 278.3, 278.4, 335.5, 353.7, 373.8, 426.7, 478.7, 487.3, 497.0, 502.7, 509.9, 519.8, 519.9, 520.2, 520.5, 523.3, 532.2, 564.0, 579.3, 591.1, 644.6, 677.5, 693.1, 697.5, 705.7, 716.5, 728.0, 739.9, 745.5, 760.8, 779.5, 830.6, 847.2, 865.1, 881.9, 921.8, 970.9, 987.6, 1007.4, 1014.1, 1040.4, 1073.8, 1115.9, 1193.7 ], "tpot_quantiles_ms": [ 55.7, 57.3, 58.4, 58.6, 59.1, 60.4, 62.3, 64.5, 65.4, 66.3, 67.1, 67.9, 68.6, 69.8, 71.3, 71.7, 73.1, 73.7, 74.6, 75.1, 76.7, 77.6, 79.4, 81.4, 82.2, 82.8, 84.3, 84.5, 86.3, 86.9, 89.1, 90.9, 92.4, 94.8, 98.6, 104.5, 106.3, 109.0, 114.1, 116.2, 119.6, 125.5, 133.9, 140.3, 147.1, 154.9, 161.4, 167.5, 170.7, 191.3, 281.6 ] }, { "x": 32.0, "output_throughput": { "median": 228.78, "min": 225.86, "max": 233.15, "n": 3 }, "median_ttft_ms": { "median": 633.41, "min": 623.24, "max": 640.21, "n": 3 }, "p99_ttft_ms": { "median": 1243.66, "min": 1236.17, "max": 1347.73, "n": 3 }, "median_tpot_ms": { "median": 106.97, "min": 103.11, "max": 109.28, "n": 3 }, "p99_tpot_ms": { "median": 261.87, "min": 244.66, "max": 267.17, "n": 3 }, "goodput_rps": { "median": 0.58, "min": 0.57, "max": 0.64, "n": 3 }, "request_throughput": { "median": 1.79, "min": 1.76, "max": 1.82, "n": 3 }, "completed": [ 128, 128, 128 ], "accept_length": { "median": 2.73, "min": 2.71, "max": 2.79, "n": 3 }, "ttft_quantiles_ms": [ 121.0, 152.4, 166.8, 318.7, 323.2, 329.0, 333.7, 334.1, 335.6, 346.6, 346.7, 346.9, 348.9, 349.1, 349.3, 349.4, 386.2, 417.9, 433.3, 469.7, 534.7, 567.8, 583.7, 605.6, 625.8, 632.0, 646.7, 664.0, 695.0, 746.6, 774.5, 791.7, 802.9, 809.8, 835.0, 852.3, 877.3, 903.7, 952.1, 1002.1, 1046.4, 1062.1, 1087.3, 1106.7, 1117.5, 1123.8, 1141.6, 1181.7, 1208.2, 1268.3, 1349.0 ], "tpot_quantiles_ms": [ 62.9, 66.1, 68.5, 69.2, 71.8, 73.0, 74.8, 76.6, 78.0, 80.6, 82.4, 84.7, 85.7, 87.8, 89.5, 91.9, 93.6, 96.0, 97.2, 98.8, 99.9, 100.7, 102.4, 103.5, 105.0, 106.6, 107.6, 109.6, 111.3, 113.9, 116.6, 118.5, 119.8, 122.8, 125.4, 129.0, 133.2, 138.1, 144.4, 153.6, 157.8, 165.8, 174.3, 177.8, 183.3, 195.7, 212.8, 222.4, 229.7, 242.9, 338.1 ] }, { "x": 64.0, "output_throughput": { "median": 276.32, "min": 268.85, "max": 289.57, "n": 3 }, "median_ttft_ms": { "median": 7772.48, "min": 7711.38, "max": 7882.33, "n": 3 }, "p99_ttft_ms": { "median": 12407.48, "min": 12146.48, "max": 12454.21, "n": 3 }, "median_tpot_ms": { "median": 138.41, "min": 134.31, "max": 140.44, "n": 3 }, "p99_tpot_ms": { "median": 362.97, "min": 331.67, "max": 409.83, "n": 3 }, "goodput_rps": { "median": 0.2, "min": 0.14, "max": 0.21, "n": 3 }, "request_throughput": { "median": 2.16, "min": 2.1, "max": 2.26, "n": 3 }, "completed": [ 256, 256, 256 ], "accept_length": { "median": 2.72, "min": 2.7, "max": 2.77, "n": 3 }, "ttft_quantiles_ms": [ 331.7, 441.0, 461.4, 464.2, 505.6, 516.0, 516.1, 525.9, 539.3, 539.7, 2396.3, 2777.0, 3136.8, 3666.8, 4127.3, 5178.1, 5951.6, 6243.3, 6476.7, 6643.8, 6851.5, 7014.2, 7247.3, 7385.8, 7588.8, 7777.6, 7951.1, 8083.3, 8171.2, 8381.1, 8546.0, 8696.1, 8851.8, 8965.1, 9119.9, 9269.9, 9429.2, 9579.9, 9707.1, 9807.5, 10053.4, 10142.7, 10354.3, 10483.4, 10772.3, 11087.2, 11383.9, 11544.5, 11825.0, 12241.0, 13786.1 ], "tpot_quantiles_ms": [ 70.9, 79.9, 82.4, 85.7, 88.6, 89.9, 92.1, 93.9, 95.8, 99.5, 101.0, 104.3, 107.0, 109.5, 112.5, 116.5, 119.1, 121.9, 125.2, 127.3, 128.8, 130.9, 132.6, 134.4, 135.8, 137.6, 139.7, 141.4, 142.9, 144.5, 146.9, 150.1, 153.9, 156.9, 163.2, 169.2, 175.1, 181.9, 188.2, 194.5, 203.3, 209.6, 220.5, 231.0, 242.2, 250.7, 261.1, 278.2, 295.5, 333.8, 479.7 ] } ] }, "fp8": { "kv_pool_tokens": 1084458, "points": [ { "x": 1.0, "output_throughput": { "median": 51.5, "min": 51.49, "max": 51.86, "n": 3 }, "median_ttft_ms": { "median": 140.78, "min": 139.21, "max": 141.28, "n": 3 }, "p99_ttft_ms": { "median": 153.97, "min": 153.21, "max": 154.42, "n": 3 }, "median_tpot_ms": { "median": 18.41, "min": 18.33, "max": 18.45, "n": 3 }, "p99_tpot_ms": { "median": 19.88, "min": 19.71, "max": 20.2, "n": 3 }, "goodput_rps": { "median": 0.4, "min": 0.4, "max": 0.41, "n": 3 }, "request_throughput": { "median": 0.4, "min": 0.4, "max": 0.41, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 44.7, 44.7, 45.1, 54.1, 64.7, 112.4, 113.8, 126.9, 129.3, 129.7, 130.6, 131.9, 133.0, 133.6, 133.7, 134.2, 135.7, 135.7, 136.7, 138.8, 139.6, 139.7, 139.7, 139.8, 140.3, 140.6, 140.7, 140.9, 141.1, 141.3, 141.4, 141.6, 142.0, 142.1, 142.1, 142.8, 143.5, 143.9, 144.3, 145.2, 145.8, 146.5, 146.8, 146.9, 148.2, 149.3, 150.2, 152.9, 153.9, 154.2, 155.7 ], "tpot_quantiles_ms": [ 17.9, 17.9, 17.9, 18.0, 18.1, 18.2, 18.2, 18.2, 18.2, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.5, 18.5, 18.5, 18.5, 18.5, 18.5, 18.5, 18.5, 18.5, 19.6, 19.9, 20.1, 20.3 ] }, { "x": 2.0, "output_throughput": { "median": 75.82, "min": 75.78, "max": 78.0, "n": 3 }, "median_ttft_ms": { "median": 74.88, "min": 74.3, "max": 75.2, "n": 3 }, "p99_ttft_ms": { "median": 77.51, "min": 77.08, "max": 92.13, "n": 3 }, "median_tpot_ms": { "median": 25.86, "min": 25.5, "max": 25.94, "n": 3 }, "p99_tpot_ms": { "median": 27.16, "min": 25.87, "max": 28.04, "n": 3 }, "goodput_rps": { "median": 0.59, "min": 0.59, "max": 0.61, "n": 3 }, "request_throughput": { "median": 0.59, "min": 0.59, "max": 0.61, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 72.3, 72.3, 72.3, 72.4, 72.5, 72.8, 72.8, 72.8, 73.0, 73.0, 73.1, 73.1, 73.5, 73.6, 73.6, 73.9, 74.2, 74.2, 74.2, 74.2, 74.3, 74.3, 74.5, 74.5, 74.6, 74.6, 74.7, 74.7, 74.7, 74.8, 74.9, 74.9, 74.9, 75.1, 75.1, 75.1, 75.3, 75.5, 75.5, 75.5, 75.7, 75.8, 75.8, 75.9, 76.2, 76.4, 76.7, 77.0, 77.1, 77.6, 95.0 ], "tpot_quantiles_ms": [ 23.9, 23.9, 23.9, 24.1, 24.3, 24.3, 24.3, 25.2, 25.2, 25.4, 25.4, 25.4, 25.5, 25.5, 25.5, 25.5, 25.6, 25.6, 25.6, 25.6, 25.7, 25.7, 25.7, 25.7, 25.7, 25.8, 25.8, 25.8, 25.8, 25.9, 25.9, 25.9, 25.9, 25.9, 25.9, 26.0, 26.0, 26.0, 26.0, 26.0, 26.0, 26.0, 26.0, 26.3, 26.4, 26.4, 26.4, 27.1, 27.2, 28.0, 28.0 ] }, { "x": 4.0, "output_throughput": { "median": 103.03, "min": 100.99, "max": 104.8, "n": 3 }, "median_ttft_ms": { "median": 84.25, "min": 83.44, "max": 84.88, "n": 3 }, "p99_ttft_ms": { "median": 85.35, "min": 84.87, "max": 94.32, "n": 3 }, "median_tpot_ms": { "median": 38.6, "min": 38.4, "max": 38.97, "n": 3 }, "p99_tpot_ms": { "median": 41.69, "min": 38.72, "max": 42.26, "n": 3 }, "goodput_rps": { "median": 0.8, "min": 0.79, "max": 0.82, "n": 3 }, "request_throughput": { "median": 0.8, "min": 0.79, "max": 0.82, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 50.5, 50.5, 50.7, 51.4, 51.6, 52.3, 52.4, 52.5, 52.6, 52.8, 53.0, 53.4, 53.6, 82.5, 82.6, 82.8, 82.9, 82.9, 83.2, 83.3, 83.4, 83.6, 83.6, 83.7, 83.8, 83.8, 83.8, 84.0, 84.2, 84.2, 84.3, 84.3, 84.4, 84.7, 84.7, 84.9, 84.9, 84.9, 84.9, 85.0, 85.0, 85.1, 85.1, 85.1, 85.1, 85.1, 85.3, 85.4, 93.9, 94.2, 94.3 ], "tpot_quantiles_ms": [ 35.3, 35.3, 35.3, 35.3, 35.6, 35.9, 35.9, 35.9, 36.1, 37.1, 37.1, 37.1, 37.4, 38.3, 38.3, 38.3, 38.3, 38.3, 38.3, 38.3, 38.4, 38.4, 38.4, 38.5, 38.5, 38.5, 38.5, 38.5, 38.5, 38.5, 38.6, 38.6, 38.7, 38.7, 38.7, 38.7, 38.7, 38.8, 38.9, 39.1, 39.1, 39.1, 39.4, 41.5, 41.5, 41.5, 41.7, 42.1, 42.1, 42.1, 42.3 ] }, { "x": 8.0, "output_throughput": { "median": 137.0, "min": 130.76, "max": 144.57, "n": 3 }, "median_ttft_ms": { "median": 125.89, "min": 113.92, "max": 129.6, "n": 3 }, "p99_ttft_ms": { "median": 673.58, "min": 662.93, "max": 929.04, "n": 3 }, "median_tpot_ms": { "median": 54.96, "min": 54.7, "max": 58.63, "n": 3 }, "p99_tpot_ms": { "median": 60.59, "min": 58.64, "max": 63.37, "n": 3 }, "goodput_rps": { "median": 1.07, "min": 1.02, "max": 1.13, "n": 3 }, "request_throughput": { "median": 1.07, "min": 1.02, "max": 1.13, "n": 3 }, "completed": [ 32, 32, 32 ], "ttft_quantiles_ms": [ 51.6, 51.9, 52.4, 54.3, 101.6, 101.7, 101.8, 102.2, 102.6, 102.7, 102.8, 102.8, 102.9, 102.9, 103.0, 103.1, 103.2, 103.3, 106.5, 106.6, 106.7, 106.7, 111.6, 111.9, 112.2, 112.5, 146.7, 156.9, 162.1, 374.8, 443.1, 443.4, 443.8, 464.9, 465.4, 506.0, 506.4, 590.0, 590.4, 627.4, 627.9, 628.1, 628.6, 662.8, 662.9, 665.1, 665.5, 673.6, 720.8, 928.9, 929.1 ], "tpot_quantiles_ms": [ 45.1, 45.1, 45.1, 45.1, 48.8, 53.6, 53.9, 53.9, 53.9, 54.2, 54.2, 54.3, 54.9, 54.9, 55.0, 55.0, 55.0, 55.0, 55.1, 55.1, 55.8, 55.8, 56.0, 56.0, 57.1, 57.1, 57.1, 57.1, 57.4, 57.5, 57.9, 57.9, 57.9, 57.9, 58.3, 58.5, 58.5, 58.6, 58.6, 58.6, 58.6, 59.1, 59.5, 59.5, 60.3, 60.3, 60.3, 60.7, 61.5, 61.5, 63.9 ] }, { "x": 16.0, "output_throughput": { "median": 189.22, "min": 188.08, "max": 195.48, "n": 3 }, "median_ttft_ms": { "median": 140.52, "min": 138.99, "max": 145.85, "n": 3 }, "p99_ttft_ms": { "median": 1136.71, "min": 1133.37, "max": 1162.21, "n": 3 }, "median_tpot_ms": { "median": 81.9, "min": 80.85, "max": 82.31, "n": 3 }, "p99_tpot_ms": { "median": 86.94, "min": 85.86, "max": 89.96, "n": 3 }, "goodput_rps": { "median": 1.1, "min": 1.04, "max": 1.15, "n": 3 }, "request_throughput": { "median": 1.48, "min": 1.47, "max": 1.53, "n": 3 }, "completed": [ 64, 64, 64 ], "ttft_quantiles_ms": [ 53.7, 56.3, 122.7, 123.6, 123.9, 124.2, 124.3, 124.7, 125.1, 125.5, 129.9, 130.8, 131.9, 133.0, 133.1, 133.3, 133.5, 133.6, 138.6, 138.6, 138.8, 138.9, 139.0, 139.1, 139.5, 140.2, 154.3, 406.3, 449.1, 589.7, 589.9, 617.6, 618.3, 654.3, 654.5, 838.8, 839.0, 1086.3, 1086.7, 1087.2, 1087.7, 1106.4, 1106.5, 1128.7, 1129.6, 1131.9, 1132.6, 1133.5, 1136.7, 1162.1, 1162.2 ], "tpot_quantiles_ms": [ 71.1, 71.1, 71.1, 71.1, 71.7, 74.9, 74.9, 75.0, 75.0, 76.5, 76.5, 76.9, 76.9, 76.9, 80.4, 80.5, 80.5, 80.9, 80.9, 80.9, 80.9, 81.2, 81.2, 81.2, 81.2, 81.5, 81.5, 82.0, 82.0, 82.0, 82.0, 82.0, 82.0, 82.3, 82.6, 82.7, 82.7, 82.7, 82.7, 82.7, 83.9, 83.9, 83.9, 83.9, 84.2, 85.5, 85.5, 86.0, 86.0, 88.1, 90.4 ] }, { "x": 32.0, "output_throughput": { "median": 283.19, "min": 281.47, "max": 286.07, "n": 3 }, "median_ttft_ms": { "median": 220.03, "min": 177.99, "max": 257.93, "n": 3 }, "p99_ttft_ms": { "median": 2141.11, "min": 2131.8, "max": 2904.21, "n": 3 }, "median_tpot_ms": { "median": 104.12, "min": 103.71, "max": 107.27, "n": 3 }, "p99_tpot_ms": { "median": 123.4, "min": 116.24, "max": 125.98, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 2.21, "min": 2.2, "max": 2.23, "n": 3 }, "completed": [ 128, 128, 128 ], "ttft_quantiles_ms": [ 56.4, 134.5, 138.3, 141.3, 155.9, 157.5, 158.7, 159.4, 159.6, 161.2, 169.7, 171.0, 172.0, 174.0, 174.2, 174.7, 175.6, 176.1, 177.0, 177.8, 178.2, 178.3, 178.5, 190.3, 220.0, 233.4, 336.4, 411.1, 576.4, 1303.1, 1304.2, 1350.2, 1354.6, 1366.8, 1371.4, 1390.9, 1391.8, 1543.2, 1544.7, 2078.9, 2081.2, 2097.2, 2098.7, 2116.0, 2131.1, 2138.8, 2140.8, 2231.7, 2235.1, 2902.4, 2904.8 ], "tpot_quantiles_ms": [ 99.3, 99.3, 100.2, 100.3, 100.3, 100.3, 102.0, 102.0, 102.6, 102.7, 102.7, 102.8, 102.8, 102.8, 102.8, 103.3, 103.6, 103.7, 103.7, 103.8, 103.8, 104.1, 104.1, 104.1, 104.1, 106.9, 106.9, 107.3, 107.3, 107.3, 107.3, 107.7, 108.1, 108.4, 108.4, 108.5, 108.5, 108.5, 108.6, 108.6, 108.6, 109.9, 109.9, 111.5, 112.2, 112.2, 115.7, 115.7, 116.2, 123.4, 128.5 ] }, { "x": 64.0, "output_throughput": { "median": 440.86, "min": 434.98, "max": 445.07, "n": 3 }, "median_ttft_ms": { "median": 318.36, "min": 305.89, "max": 323.71, "n": 3 }, "p99_ttft_ms": { "median": 4068.67, "min": 4068.67, "max": 4490.2, "n": 3 }, "median_tpot_ms": { "median": 131.78, "min": 129.56, "max": 132.46, "n": 3 }, "p99_tpot_ms": { "median": 158.33, "min": 158.26, "max": 161.5, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 3.44, "min": 3.4, "max": 3.48, "n": 3 }, "completed": [ 256, 256, 256 ], "ttft_quantiles_ms": [ 62.8, 147.8, 152.2, 167.4, 170.0, 171.1, 173.8, 181.5, 198.6, 198.8, 203.9, 204.4, 212.6, 214.0, 247.4, 253.5, 263.1, 274.1, 293.0, 300.3, 305.9, 315.0, 315.2, 316.6, 318.4, 321.5, 509.7, 691.1, 1316.4, 1366.2, 1493.2, 1552.3, 1555.3, 1665.8, 1719.7, 2289.2, 2361.6, 2465.2, 2522.2, 2633.0, 2681.2, 3256.6, 3343.1, 3448.5, 3508.6, 3610.3, 3658.1, 3744.6, 4050.1, 4068.6, 4490.3 ], "tpot_quantiles_ms": [ 124.9, 125.4, 126.0, 126.4, 126.6, 126.6, 126.8, 126.8, 126.8, 126.9, 127.8, 128.0, 129.1, 129.5, 129.6, 129.6, 129.6, 129.9, 130.3, 130.3, 130.4, 130.6, 130.6, 130.8, 130.9, 130.9, 130.9, 131.8, 132.3, 134.0, 134.1, 135.1, 135.4, 135.7, 136.6, 138.4, 141.8, 142.8, 143.1, 143.3, 144.4, 146.2, 149.5, 149.5, 150.5, 150.7, 152.3, 153.8, 153.8, 158.3, 164.8 ] } ] }, "no-cuda-graph": { "kv_pool_tokens": 235589, "points": [ { "x": 1.0, "output_throughput": { "median": 28.72, "min": 28.27, "max": 29.06, "n": 3 }, "median_ttft_ms": { "median": 291.75, "min": 291.52, "max": 296.67, "n": 3 }, "p99_ttft_ms": { "median": 321.82, "min": 321.8, "max": 328.03, "n": 3 }, "median_tpot_ms": { "median": 33.09, "min": 32.44, "max": 33.43, "n": 3 }, "p99_tpot_ms": { "median": 33.16, "min": 32.52, "max": 33.81, "n": 3 }, "goodput_rps": { "median": 0.22, "min": 0.22, "max": 0.23, "n": 3 }, "request_throughput": { "median": 0.22, "min": 0.22, "max": 0.23, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 61.6, 61.6, 62.5, 70.6, 102.5, 225.2, 228.4, 265.0, 267.3, 274.5, 276.4, 277.5, 279.2, 279.7, 280.0, 280.1, 284.0, 284.0, 287.2, 288.4, 288.5, 288.9, 290.0, 291.8, 293.5, 293.7, 294.6, 294.9, 294.9, 295.2, 295.5, 298.2, 298.7, 299.1, 299.1, 299.2, 300.5, 301.2, 301.9, 304.6, 306.5, 307.6, 307.7, 308.2, 312.2, 312.6, 318.0, 319.7, 323.4, 323.5, 329.5 ], "tpot_quantiles_ms": [ 32.4, 32.4, 32.4, 32.4, 32.4, 32.4, 32.4, 32.4, 32.4, 32.4, 32.5, 32.5, 32.5, 32.5, 32.5, 32.5, 32.5, 32.5, 32.7, 32.8, 32.8, 32.9, 33.0, 33.0, 33.0, 33.1, 33.1, 33.1, 33.1, 33.1, 33.1, 33.1, 33.1, 33.1, 33.1, 33.2, 33.2, 33.2, 33.3, 33.4, 33.4, 33.4, 33.4, 33.4, 33.5, 33.5, 33.6, 33.7, 33.8, 33.8, 33.8 ] }, { "x": 2.0, "output_throughput": { "median": 43.15, "min": 43.02, "max": 44.11, "n": 3 }, "median_ttft_ms": { "median": 80.08, "min": 77.39, "max": 80.11, "n": 3 }, "p99_ttft_ms": { "median": 98.58, "min": 97.45, "max": 99.68, "n": 3 }, "median_tpot_ms": { "median": 46.18, "min": 45.68, "max": 46.32, "n": 3 }, "p99_tpot_ms": { "median": 47.54, "min": 46.87, "max": 47.67, "n": 3 }, "goodput_rps": { "median": 0.34, "min": 0.34, "max": 0.34, "n": 3 }, "request_throughput": { "median": 0.34, "min": 0.34, "max": 0.34, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 59.9, 59.9, 59.9, 60.3, 60.6, 60.9, 61.2, 62.2, 62.3, 62.7, 62.7, 63.0, 63.0, 63.3, 63.4, 63.4, 63.5, 63.5, 63.8, 64.5, 64.8, 64.9, 65.4, 66.3, 66.3, 66.8, 91.0, 91.2, 91.2, 91.8, 92.3, 92.5, 93.0, 93.4, 93.4, 93.9, 94.2, 94.7, 95.0, 95.1, 95.9, 96.0, 96.2, 96.3, 96.3, 96.4, 97.6, 97.9, 98.0, 98.7, 100.9 ], "tpot_quantiles_ms": [ 41.9, 41.9, 42.1, 42.5, 42.8, 43.1, 43.4, 45.2, 45.2, 45.2, 45.3, 45.4, 45.5, 45.5, 45.5, 45.6, 45.6, 45.6, 45.6, 45.8, 45.8, 45.8, 45.9, 46.0, 46.0, 46.1, 46.1, 46.2, 46.2, 46.2, 46.3, 46.4, 46.4, 46.4, 46.4, 46.5, 46.5, 46.5, 46.7, 46.7, 46.7, 46.7, 46.7, 46.8, 46.9, 47.0, 47.0, 47.3, 47.5, 47.6, 47.7 ] }, { "x": 4.0, "output_throughput": { "median": 56.2, "min": 55.71, "max": 56.23, "n": 3 }, "median_ttft_ms": { "median": 121.78, "min": 119.91, "max": 123.16, "n": 3 }, "p99_ttft_ms": { "median": 125.65, "min": 125.29, "max": 322.14, "n": 3 }, "median_tpot_ms": { "median": 71.49, "min": 71.37, "max": 71.64, "n": 3 }, "p99_tpot_ms": { "median": 74.53, "min": 72.19, "max": 75.02, "n": 3 }, "goodput_rps": { "median": 0.44, "min": 0.44, "max": 0.44, "n": 3 }, "request_throughput": { "median": 0.44, "min": 0.44, "max": 0.44, "n": 3 }, "completed": [ 16, 16, 16 ], "ttft_quantiles_ms": [ 61.6, 61.6, 62.0, 63.8, 65.1, 65.3, 65.6, 66.6, 67.2, 67.2, 68.4, 68.7, 119.0, 119.0, 119.1, 119.8, 119.9, 119.9, 120.0, 120.8, 121.0, 121.1, 121.1, 121.3, 121.4, 121.4, 121.4, 121.5, 121.8, 121.8, 121.9, 122.6, 122.9, 123.0, 123.0, 123.1, 123.3, 123.4, 124.6, 124.9, 125.0, 125.0, 125.2, 125.3, 125.5, 125.6, 125.7, 261.5, 322.0, 322.1, 322.1 ], "tpot_quantiles_ms": [ 65.8, 65.8, 65.8, 65.8, 66.3, 67.0, 67.0, 67.0, 67.5, 68.5, 68.5, 68.5, 68.9, 70.4, 70.4, 70.4, 70.8, 70.8, 70.9, 70.9, 70.9, 71.1, 71.1, 71.1, 71.4, 71.4, 71.4, 71.4, 71.5, 71.6, 71.6, 71.6, 71.8, 71.8, 71.8, 71.8, 71.8, 72.0, 72.2, 72.5, 72.5, 72.5, 72.9, 74.1, 74.1, 74.1, 74.6, 74.6, 74.6, 74.6, 75.1 ] }, { "x": 8.0, "output_throughput": { "median": 73.46, "min": 70.35, "max": 76.17, "n": 3 }, "median_ttft_ms": { "median": 233.0, "min": 205.99, "max": 252.63, "n": 3 }, "p99_ttft_ms": { "median": 1051.1, "min": 998.78, "max": 1071.99, "n": 3 }, "median_tpot_ms": { "median": 104.77, "min": 103.05, "max": 111.87, "n": 3 }, "p99_tpot_ms": { "median": 115.14, "min": 111.69, "max": 117.26, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.57, "min": 0.55, "max": 0.6, "n": 3 }, "completed": [ 32, 32, 32 ], "ttft_quantiles_ms": [ 64.7, 66.6, 69.5, 87.6, 161.0, 161.3, 161.9, 162.1, 162.3, 162.6, 162.8, 166.0, 166.6, 167.1, 167.6, 173.1, 173.1, 173.2, 178.7, 178.9, 179.2, 179.4, 199.9, 200.2, 200.4, 200.7, 292.7, 307.7, 370.9, 914.4, 915.2, 915.3, 915.7, 968.3, 968.9, 969.6, 998.0, 998.2, 998.5, 998.8, 1003.8, 1004.0, 1004.7, 1005.4, 1050.0, 1050.6, 1051.0, 1070.8, 1071.2, 1071.8, 1072.1 ], "tpot_quantiles_ms": [ 85.2, 85.2, 85.2, 85.2, 90.8, 101.5, 101.6, 101.6, 101.6, 101.6, 101.6, 101.6, 102.1, 102.1, 102.1, 102.1, 103.0, 103.1, 103.1, 103.1, 106.3, 106.3, 106.3, 106.3, 106.5, 106.5, 106.5, 106.5, 108.4, 108.4, 108.4, 108.4, 108.4, 110.1, 110.1, 110.1, 110.1, 111.7, 112.0, 112.0, 112.0, 112.0, 112.9, 113.3, 113.3, 113.3, 114.6, 114.6, 114.6, 114.6, 119.0 ] }, { "x": 16.0, "output_throughput": { "median": 100.89, "min": 99.05, "max": 101.12, "n": 3 }, "median_ttft_ms": { "median": 251.96, "min": 246.66, "max": 279.06, "n": 3 }, "p99_ttft_ms": { "median": 1702.44, "min": 1516.7, "max": 1991.94, "n": 3 }, "median_tpot_ms": { "median": 157.39, "min": 156.62, "max": 159.1, "n": 3 }, "p99_tpot_ms": { "median": 164.51, "min": 163.52, "max": 165.03, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.79, "min": 0.77, "max": 0.79, "n": 3 }, "completed": [ 64, 64, 64 ], "ttft_quantiles_ms": [ 68.3, 69.2, 200.9, 201.0, 201.3, 201.5, 208.0, 208.2, 208.5, 208.6, 211.8, 212.7, 213.5, 221.9, 222.1, 222.3, 222.5, 226.5, 226.8, 226.9, 227.0, 244.1, 244.3, 244.4, 244.6, 266.4, 295.2, 1224.1, 1224.6, 1226.0, 1323.3, 1324.5, 1325.7, 1492.5, 1492.7, 1493.0, 1493.9, 1494.1, 1494.2, 1494.3, 1516.1, 1516.3, 1516.5, 1516.7, 1541.4, 1542.0, 1542.7, 1543.6, 1702.4, 1731.9, 1992.0 ], "tpot_quantiles_ms": [ 139.4, 139.4, 139.4, 139.4, 140.5, 143.9, 143.9, 143.9, 143.9, 147.6, 148.5, 151.4, 151.4, 151.4, 151.8, 151.8, 151.8, 152.1, 152.1, 152.1, 152.1, 154.5, 154.5, 154.5, 154.5, 156.6, 156.6, 156.6, 156.6, 157.1, 157.1, 157.1, 157.4, 157.8, 159.1, 159.1, 159.1, 159.1, 159.5, 159.5, 159.5, 159.5, 160.9, 161.3, 161.3, 161.3, 162.7, 163.1, 163.1, 163.1, 169.9 ] }, { "x": 32.0, "output_throughput": { "median": 148.92, "min": 148.7, "max": 150.36, "n": 3 }, "median_ttft_ms": { "median": 395.26, "min": 380.42, "max": 537.5, "n": 3 }, "p99_ttft_ms": { "median": 3189.57, "min": 3058.96, "max": 3219.1, "n": 3 }, "median_tpot_ms": { "median": 203.06, "min": 201.41, "max": 205.85, "n": 3 }, "p99_tpot_ms": { "median": 222.1, "min": 215.66, "max": 225.47, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.16, "min": 1.16, "max": 1.17, "n": 3 }, "completed": [ 128, 128, 128 ], "ttft_quantiles_ms": [ 72.1, 188.6, 204.1, 207.6, 209.9, 232.7, 234.0, 277.0, 278.3, 279.4, 286.1, 348.0, 349.5, 350.9, 355.0, 360.0, 361.6, 376.5, 377.8, 378.7, 380.9, 394.8, 521.2, 536.8, 537.0, 537.2, 1237.9, 1241.1, 1258.8, 1319.7, 1323.4, 1360.2, 1592.5, 1595.0, 1618.3, 1619.4, 2578.2, 2580.9, 2620.7, 2622.0, 2705.6, 2707.3, 2751.3, 2753.2, 2876.0, 2876.1, 2909.9, 2911.5, 3058.9, 3174.2, 3235.5 ], "tpot_quantiles_ms": [ 193.3, 195.8, 197.1, 197.1, 198.1, 198.1, 198.1, 198.1, 198.4, 198.4, 199.4, 200.0, 200.0, 200.0, 200.1, 200.1, 200.1, 200.2, 200.2, 201.4, 201.5, 202.6, 202.6, 203.0, 203.1, 203.8, 203.8, 205.8, 205.8, 205.9, 207.2, 208.0, 208.0, 208.3, 208.3, 210.6, 210.6, 210.8, 211.3, 212.8, 212.8, 214.0, 214.0, 214.1, 214.7, 214.7, 215.2, 221.0, 221.0, 225.5, 233.1 ] }, { "x": 64.0, "output_throughput": { "median": 235.79, "min": 233.14, "max": 237.97, "n": 3 }, "median_ttft_ms": { "median": 564.89, "min": 554.64, "max": 587.32, "n": 3 }, "p99_ttft_ms": { "median": 5769.61, "min": 5706.99, "max": 5886.08, "n": 3 }, "median_tpot_ms": { "median": 253.77, "min": 250.18, "max": 255.08, "n": 3 }, "p99_tpot_ms": { "median": 283.65, "min": 278.33, "max": 290.64, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.84, "min": 1.82, "max": 1.86, "n": 3 }, "completed": [ 256, 256, 256 ], "ttft_quantiles_ms": [ 83.3, 213.5, 216.7, 287.3, 289.2, 289.5, 293.2, 308.4, 380.4, 384.7, 400.7, 403.9, 405.6, 410.5, 457.3, 459.5, 483.0, 483.1, 483.5, 483.9, 548.3, 553.4, 559.7, 564.6, 571.8, 586.7, 1166.7, 1645.6, 1656.9, 1671.3, 1679.8, 1698.6, 2524.8, 2833.3, 3018.4, 3052.1, 3057.1, 3065.8, 3876.0, 4232.2, 4407.6, 4419.9, 4440.8, 4443.5, 5250.8, 5444.8, 5527.7, 5609.9, 5675.4, 5706.9, 5962.3 ], "tpot_quantiles_ms": [ 242.6, 243.6, 243.6, 244.9, 245.4, 245.7, 246.3, 246.6, 247.2, 247.2, 247.6, 248.6, 248.7, 248.8, 248.8, 248.9, 249.7, 250.0, 250.2, 250.2, 251.0, 251.7, 252.2, 252.2, 252.3, 253.6, 253.6, 253.7, 253.8, 254.7, 255.1, 255.1, 256.3, 256.7, 260.1, 262.0, 264.4, 265.6, 267.3, 267.5, 268.9, 272.9, 275.4, 276.3, 276.3, 278.5, 279.8, 283.6, 286.3, 290.6, 301.5 ] } ] }, "default-kv-max": { "kv_pool_tokens": 383065, "points": [ { "x": 1.0, "output_throughput": { "median": 29.22, "min": 28.89, "max": 29.42, "n": 3 }, "median_ttft_ms": { "median": 272.86, "min": 269.53, "max": 309.68, "n": 3 }, "p99_ttft_ms": { "median": 306.17, "min": 298.7, "max": 455.16, "n": 3 }, "median_tpot_ms": { "median": 32.12, "min": 32.08, "max": 32.62, "n": 3 }, "p99_tpot_ms": { "median": 34.1, "min": 32.21, "max": 34.88, "n": 3 }, "p99_itl_ms": { "median": 40.89, "min": 40.71, "max": 42.3, "n": 3 }, "median_itl_ms": { "median": 32.03, "min": 32.03, "max": 32.55, "n": 3 }, "goodput_rps": { "median": 0.23, "min": 0.23, "max": 0.23, "n": 3 }, "request_throughput": { "median": 0.23, "min": 0.23, "max": 0.23, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.0, 42.0, 42.1, 55.8, 92.2, 208.7, 247.8, 249.7, 252.0, 253.2, 257.2, 257.7, 260.0, 264.0, 265.5, 268.8, 269.4, 269.4, 269.6, 270.5, 270.6, 272.1, 272.6, 275.1, 275.1, 275.8, 276.4, 277.0, 277.8, 279.3, 282.7, 283.2, 286.7, 287.4, 287.4, 287.6, 297.1, 298.8, 300.8, 301.6, 307.5, 309.0, 309.6, 309.7, 310.6, 317.5, 337.3, 353.2, 374.0, 376.8, 469.0 ], "tpot_quantiles_ms": [ 31.8, 31.8, 31.8, 31.9, 31.9, 31.9, 32.0, 32.0, 32.0, 32.0, 32.0, 32.0, 32.0, 32.0, 32.0, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.2, 32.2, 32.2, 32.2, 32.2, 32.2, 32.2, 32.2, 32.3, 32.3, 32.3, 32.4, 32.5, 32.5, 32.5, 32.5, 32.7, 32.8, 32.9, 32.9, 32.9, 33.0, 33.4, 34.2, 34.6, 34.9 ] }, { "x": 2.0, "output_throughput": { "median": 44.02, "min": 43.98, "max": 45.07, "n": 3 }, "median_ttft_ms": { "median": 69.58, "min": 69.22, "max": 71.23, "n": 3 }, "p99_ttft_ms": { "median": 82.91, "min": 76.05, "max": 89.07, "n": 3 }, "median_tpot_ms": { "median": 44.77, "min": 44.72, "max": 45.22, "n": 3 }, "p99_tpot_ms": { "median": 46.13, "min": 45.75, "max": 48.54, "n": 3 }, "p99_itl_ms": { "median": 57.74, "min": 55.45, "max": 58.16, "n": 3 }, "median_itl_ms": { "median": 45.11, "min": 44.62, "max": 45.42, "n": 3 }, "goodput_rps": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "request_throughput": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 60.2, 60.2, 60.4, 60.8, 60.9, 61.0, 61.2, 61.3, 61.6, 62.3, 62.4, 62.7, 62.8, 62.8, 62.8, 63.1, 63.3, 63.3, 63.4, 63.5, 63.6, 63.7, 64.1, 65.2, 66.7, 68.6, 72.5, 73.3, 73.5, 73.5, 73.6, 73.9, 73.9, 73.9, 73.9, 74.0, 74.0, 74.1, 74.2, 74.2, 74.4, 74.8, 75.0, 75.6, 75.6, 75.7, 75.8, 76.2, 80.1, 83.4, 91.4 ], "tpot_quantiles_ms": [ 41.3, 41.3, 41.4, 41.8, 42.0, 44.0, 44.1, 44.2, 44.2, 44.2, 44.2, 44.3, 44.3, 44.4, 44.4, 44.5, 44.5, 44.5, 44.6, 44.7, 44.8, 44.8, 44.9, 44.9, 44.9, 44.9, 45.0, 45.0, 45.0, 45.1, 45.1, 45.1, 45.1, 45.2, 45.2, 45.2, 45.3, 45.5, 45.6, 45.7, 45.8, 45.8, 45.8, 45.9, 45.9, 45.9, 46.0, 46.0, 46.1, 48.5, 48.6 ] }, { "x": 4.0, "output_throughput": { "median": 57.28, "min": 56.71, "max": 57.8, "n": 3 }, "median_ttft_ms": { "median": 113.88, "min": 113.03, "max": 114.57, "n": 3 }, "p99_ttft_ms": { "median": 116.3, "min": 115.71, "max": 118.19, "n": 3 }, "median_tpot_ms": { "median": 70.08, "min": 69.83, "max": 70.54, "n": 3 }, "p99_tpot_ms": { "median": 73.23, "min": 70.53, "max": 74.22, "n": 3 }, "p99_itl_ms": { "median": 80.49, "min": 76.86, "max": 88.03, "n": 3 }, "median_itl_ms": { "median": 69.64, "min": 69.24, "max": 69.93, "n": 3 }, "goodput_rps": { "median": 0.45, "min": 0.44, "max": 0.45, "n": 3 }, "request_throughput": { "median": 0.45, "min": 0.44, "max": 0.45, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 45.9, 45.9, 46.4, 47.2, 47.5, 47.6, 47.7, 47.7, 47.8, 48.6, 49.2, 49.6, 52.0, 110.1, 110.3, 110.4, 111.8, 111.8, 111.9, 111.9, 113.0, 113.1, 113.1, 113.1, 113.1, 113.2, 113.8, 114.0, 114.1, 114.4, 114.7, 114.8, 114.9, 115.0, 115.0, 115.1, 115.1, 115.1, 115.2, 115.3, 115.5, 115.7, 116.1, 116.3, 116.3, 116.5, 116.6, 116.6, 117.8, 118.0, 118.2 ], "tpot_quantiles_ms": [ 64.4, 64.4, 64.4, 64.4, 64.9, 65.7, 65.7, 65.7, 66.2, 67.2, 67.2, 67.2, 67.7, 68.4, 68.4, 68.4, 68.9, 68.9, 69.5, 69.5, 69.5, 69.7, 69.7, 69.7, 69.8, 69.8, 69.8, 70.0, 70.0, 70.0, 70.0, 70.2, 70.2, 70.2, 70.2, 70.2, 70.4, 70.6, 70.7, 72.2, 72.2, 72.2, 72.7, 72.8, 72.8, 72.8, 73.3, 73.8, 73.8, 73.8, 74.3 ] }, { "x": 8.0, "output_throughput": { "median": 73.79, "min": 71.01, "max": 78.23, "n": 3 }, "median_ttft_ms": { "median": 211.08, "min": 185.31, "max": 222.65, "n": 3 }, "p99_ttft_ms": { "median": 1042.79, "min": 1029.43, "max": 1373.7, "n": 3 }, "median_tpot_ms": { "median": 102.42, "min": 102.33, "max": 109.79, "n": 3 }, "p99_tpot_ms": { "median": 113.5, "min": 108.08, "max": 117.16, "n": 3 }, "p99_itl_ms": { "median": 133.63, "min": 118.19, "max": 149.11, "n": 3 }, "median_itl_ms": { "median": 102.01, "min": 100.52, "max": 107.84, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.58, "min": 0.55, "max": 0.61, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 47.4, 48.9, 50.1, 51.5, 149.0, 149.2, 149.4, 149.7, 149.8, 150.0, 150.1, 150.2, 150.5, 150.6, 150.9, 160.6, 160.7, 160.8, 160.9, 160.9, 161.0, 161.1, 162.8, 162.9, 163.0, 163.0, 261.3, 296.2, 398.8, 592.2, 894.0, 894.2, 964.8, 965.2, 965.9, 966.7, 1024.0, 1024.1, 1024.2, 1028.0, 1028.5, 1028.7, 1029.5, 1041.6, 1042.2, 1042.7, 1134.6, 1135.0, 1135.5, 1373.3, 1373.8 ], "tpot_quantiles_ms": [ 82.7, 82.7, 82.7, 82.7, 88.0, 99.0, 99.0, 100.5, 100.5, 100.5, 100.8, 100.8, 100.8, 101.3, 102.1, 102.1, 102.1, 102.8, 102.8, 103.3, 103.3, 103.3, 103.3, 103.8, 103.8, 103.8, 105.1, 106.7, 107.3, 107.3, 107.3, 107.3, 108.0, 108.5, 108.6, 108.6, 108.7, 109.4, 109.8, 109.8, 109.8, 109.8, 110.7, 112.9, 112.9, 112.9, 113.2, 113.2, 113.2, 113.2, 118.9 ] }, { "x": 16.0, "output_throughput": { "median": 101.02, "min": 100.22, "max": 105.18, "n": 3 }, "median_ttft_ms": { "median": 261.45, "min": 235.99, "max": 280.21, "n": 3 }, "p99_ttft_ms": { "median": 1726.44, "min": 1619.39, "max": 2325.49, "n": 3 }, "median_tpot_ms": { "median": 153.4, "min": 152.08, "max": 159.12, "n": 3 }, "p99_tpot_ms": { "median": 160.81, "min": 157.75, "max": 166.09, "n": 3 }, "p99_itl_ms": { "median": 189.39, "min": 167.62, "max": 199.9, "n": 3 }, "median_itl_ms": { "median": 151.57, "min": 147.87, "max": 152.48, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.79, "min": 0.78, "max": 0.82, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.011 }, "ttft_quantiles_ms": [ 50.0, 51.3, 168.4, 168.7, 171.3, 172.2, 172.6, 185.3, 185.5, 185.7, 185.9, 193.4, 193.5, 193.7, 201.6, 202.5, 203.1, 203.8, 204.8, 204.9, 205.1, 205.3, 245.4, 258.3, 262.3, 262.4, 298.0, 613.3, 657.4, 848.4, 1434.1, 1434.5, 1434.6, 1434.9, 1617.8, 1618.1, 1618.4, 1619.3, 1654.2, 1654.3, 1654.6, 1671.2, 1672.0, 1673.1, 1725.1, 1725.5, 1726.2, 2322.3, 2323.2, 2324.2, 2325.8 ], "tpot_quantiles_ms": [ 134.5, 134.5, 135.1, 135.1, 136.0, 144.4, 144.4, 144.4, 144.4, 145.0, 145.0, 145.0, 145.0, 147.0, 147.0, 147.2, 147.2, 147.3, 147.3, 150.8, 150.8, 150.9, 152.1, 152.1, 152.1, 152.1, 153.4, 153.4, 153.4, 153.4, 153.4, 154.0, 154.0, 154.6, 154.6, 154.6, 154.6, 154.7, 154.7, 155.7, 158.7, 159.5, 159.5, 159.5, 159.5, 159.8, 159.8, 159.8, 159.8, 161.3, 168.5 ] }, { "x": 32.0, "output_throughput": { "median": 151.77, "min": 149.23, "max": 152.34, "n": 3 }, "median_ttft_ms": { "median": 432.09, "min": 422.06, "max": 467.18, "n": 3 }, "p99_ttft_ms": { "median": 3207.42, "min": 3023.16, "max": 3961.27, "n": 3 }, "median_tpot_ms": { "median": 198.65, "min": 197.67, "max": 206.74, "n": 3 }, "p99_tpot_ms": { "median": 225.87, "min": 214.95, "max": 227.44, "n": 3 }, "p99_itl_ms": { "median": 270.5, "min": 248.36, "max": 301.25, "n": 3 }, "median_itl_ms": { "median": 196.29, "min": 194.56, "max": 200.58, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.19, "min": 1.17, "max": 1.19, "n": 3 }, "completed": [ 128, 128, 128 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 85.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 51.6, 170.0, 173.9, 178.3, 210.2, 214.1, 214.4, 215.5, 241.1, 241.6, 242.3, 264.7, 270.6, 278.6, 335.4, 346.3, 346.6, 346.9, 361.4, 361.6, 409.2, 418.3, 419.4, 421.8, 432.1, 444.5, 586.1, 665.6, 1874.6, 1879.0, 1887.8, 1889.2, 1925.6, 1926.8, 1934.0, 1935.7, 2244.9, 2246.3, 2826.5, 2827.5, 2828.8, 2914.4, 2927.6, 2929.8, 3008.1, 3009.3, 3022.3, 3207.1, 3207.3, 3960.2, 3961.4 ], "tpot_quantiles_ms": [ 187.6, 187.6, 193.1, 193.7, 194.1, 194.1, 194.8, 194.8, 195.2, 195.2, 196.4, 196.4, 196.8, 196.8, 197.1, 197.1, 197.1, 197.7, 197.8, 197.9, 198.1, 198.1, 198.4, 198.7, 199.1, 199.1, 202.3, 202.3, 202.5, 202.6, 205.4, 205.4, 206.7, 206.7, 206.7, 207.4, 207.4, 207.4, 208.9, 209.7, 211.2, 211.2, 212.4, 213.1, 213.1, 214.2, 214.2, 217.5, 217.5, 217.5, 230.8 ] }, { "x": 64.0, "output_throughput": { "median": 238.72, "min": 237.93, "max": 241.79, "n": 3 }, "median_ttft_ms": { "median": 588.14, "min": 571.73, "max": 594.15, "n": 3 }, "p99_ttft_ms": { "median": 5491.19, "min": 5470.34, "max": 5572.74, "n": 3 }, "median_tpot_ms": { "median": 252.57, "min": 246.82, "max": 253.4, "n": 3 }, "p99_tpot_ms": { "median": 282.9, "min": 279.78, "max": 286.38, "n": 3 }, "p99_itl_ms": { "median": 331.8, "min": 321.83, "max": 385.97, "n": 3 }, "median_itl_ms": { "median": 243.02, "min": 240.49, "max": 245.36, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.87, "min": 1.86, "max": 1.89, "n": 3 }, "completed": [ 256, 256, 256 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 56.0, 188.9, 212.7, 245.6, 270.4, 281.9, 282.2, 285.9, 286.5, 301.6, 375.5, 416.5, 419.6, 427.2, 484.2, 489.0, 494.8, 504.3, 510.3, 516.9, 557.2, 558.8, 571.7, 587.9, 593.0, 593.9, 625.1, 1803.7, 1809.8, 1831.2, 1879.3, 1900.5, 2139.5, 3066.9, 3082.4, 3089.0, 3094.2, 3158.1, 3188.3, 3415.2, 4327.6, 4348.0, 4363.8, 4420.9, 4457.7, 4692.0, 5047.4, 5468.6, 5490.0, 5490.9, 5573.0 ], "tpot_quantiles_ms": [ 239.1, 240.9, 240.9, 240.9, 242.5, 243.4, 243.4, 243.6, 244.8, 244.9, 244.9, 245.5, 245.5, 245.5, 246.0, 246.2, 246.3, 247.3, 247.5, 247.5, 247.5, 249.0, 249.6, 250.2, 250.3, 251.6, 251.8, 252.3, 252.6, 252.9, 253.4, 253.4, 254.6, 254.6, 256.1, 258.0, 258.8, 259.6, 259.6, 261.8, 262.5, 266.2, 267.9, 268.8, 269.7, 271.8, 272.5, 278.0, 278.8, 278.8, 291.3 ] } ] }, "fp8-kv-max": { "kv_pool_tokens": 1378545, "points": [ { "x": 1.0, "output_throughput": { "median": 51.63, "min": 51.41, "max": 51.65, "n": 3 }, "median_ttft_ms": { "median": 141.33, "min": 138.0, "max": 141.44, "n": 3 }, "p99_ttft_ms": { "median": 169.25, "min": 165.7, "max": 181.61, "n": 3 }, "median_tpot_ms": { "median": 18.31, "min": 18.29, "max": 18.34, "n": 3 }, "p99_tpot_ms": { "median": 20.87, "min": 20.43, "max": 21.26, "n": 3 }, "p99_itl_ms": { "median": 24.4, "min": 23.95, "max": 24.66, "n": 3 }, "median_itl_ms": { "median": 18.23, "min": 18.22, "max": 18.26, "n": 3 }, "goodput_rps": { "median": 0.4, "min": 0.4, "max": 0.4, "n": 3 }, "request_throughput": { "median": 0.4, "min": 0.4, "max": 0.4, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.0, 41.0, 41.5, 48.8, 63.0, 110.1, 112.8, 126.5, 127.4, 129.6, 130.1, 132.4, 132.9, 133.8, 134.1, 134.3, 136.4, 136.4, 137.7, 138.2, 138.3, 138.3, 138.5, 139.4, 139.4, 139.7, 140.5, 140.8, 141.0, 141.7, 141.8, 142.0, 142.1, 143.2, 143.2, 143.7, 143.7, 144.8, 145.0, 145.3, 146.7, 149.0, 150.0, 152.9, 153.2, 153.3, 153.6, 154.0, 167.9, 172.0, 186.6 ], "tpot_quantiles_ms": [ 17.8, 17.8, 17.9, 17.9, 17.9, 17.9, 17.9, 17.9, 18.0, 18.0, 18.0, 18.0, 18.1, 18.1, 18.1, 18.1, 18.2, 18.2, 18.2, 18.2, 18.2, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.3, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.4, 18.7, 18.8, 20.1, 20.6, 20.7, 20.9, 21.5 ] }, { "x": 2.0, "output_throughput": { "median": 76.62, "min": 74.99, "max": 77.58, "n": 3 }, "median_ttft_ms": { "median": 70.21, "min": 69.75, "max": 71.14, "n": 3 }, "p99_ttft_ms": { "median": 73.13, "min": 72.13, "max": 78.45, "n": 3 }, "median_tpot_ms": { "median": 25.68, "min": 25.64, "max": 25.97, "n": 3 }, "p99_tpot_ms": { "median": 26.37, "min": 26.2, "max": 29.6, "n": 3 }, "p99_itl_ms": { "median": 27.78, "min": 27.42, "max": 33.15, "n": 3 }, "median_itl_ms": { "median": 25.79, "min": 25.68, "max": 25.93, "n": 3 }, "goodput_rps": { "median": 0.6, "min": 0.59, "max": 0.61, "n": 3 }, "request_throughput": { "median": 0.6, "min": 0.59, "max": 0.61, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 66.4, 66.4, 67.9, 68.0, 68.1, 68.3, 68.4, 68.7, 68.7, 68.8, 68.8, 69.0, 69.0, 69.3, 69.3, 69.3, 69.6, 69.6, 69.7, 69.8, 70.0, 70.0, 70.2, 70.2, 70.2, 70.4, 70.4, 70.5, 70.5, 70.6, 70.6, 70.7, 70.7, 70.8, 70.8, 70.8, 71.0, 71.3, 71.5, 71.5, 71.6, 71.7, 71.8, 71.8, 72.2, 72.5, 72.6, 72.6, 73.0, 73.2, 79.7 ], "tpot_quantiles_ms": [ 23.9, 23.9, 23.9, 24.2, 24.3, 25.2, 25.2, 25.2, 25.3, 25.4, 25.4, 25.4, 25.4, 25.5, 25.6, 25.6, 25.6, 25.6, 25.6, 25.6, 25.6, 25.6, 25.7, 25.7, 25.7, 25.7, 25.7, 25.7, 25.8, 25.8, 25.9, 25.9, 26.0, 26.0, 26.0, 26.0, 26.0, 26.0, 26.1, 26.2, 26.2, 26.2, 26.2, 26.3, 26.3, 26.3, 26.3, 26.4, 26.4, 29.6, 29.6 ] }, { "x": 4.0, "output_throughput": { "median": 103.81, "min": 101.42, "max": 104.92, "n": 3 }, "median_ttft_ms": { "median": 80.87, "min": 80.74, "max": 81.5, "n": 3 }, "p99_ttft_ms": { "median": 82.65, "min": 82.06, "max": 89.27, "n": 3 }, "median_tpot_ms": { "median": 38.34, "min": 38.27, "max": 38.48, "n": 3 }, "p99_tpot_ms": { "median": 39.5, "min": 38.7, "max": 43.5, "n": 3 }, "p99_itl_ms": { "median": 41.46, "min": 40.36, "max": 49.45, "n": 3 }, "median_itl_ms": { "median": 38.12, "min": 37.95, "max": 38.44, "n": 3 }, "goodput_rps": { "median": 0.81, "min": 0.79, "max": 0.82, "n": 3 }, "request_throughput": { "median": 0.81, "min": 0.79, "max": 0.82, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 47.5, 47.5, 47.5, 47.8, 48.6, 48.7, 48.9, 49.0, 49.1, 49.2, 49.2, 49.7, 50.6, 80.3, 80.4, 80.6, 80.6, 80.6, 80.7, 80.7, 80.8, 80.8, 80.8, 80.9, 80.9, 81.0, 81.0, 81.0, 81.1, 81.2, 81.3, 81.4, 81.4, 81.5, 81.5, 81.6, 81.7, 81.8, 81.9, 82.0, 82.0, 82.1, 82.5, 82.6, 82.6, 82.7, 82.8, 82.8, 89.1, 89.2, 89.3 ], "tpot_quantiles_ms": [ 36.1, 36.1, 36.1, 36.1, 36.3, 36.3, 36.3, 36.4, 36.6, 37.1, 37.1, 37.1, 37.3, 37.4, 37.4, 37.4, 37.7, 37.7, 38.1, 38.1, 38.1, 38.1, 38.1, 38.2, 38.3, 38.3, 38.3, 38.4, 38.4, 38.5, 38.5, 38.5, 38.5, 38.5, 38.5, 38.5, 38.6, 38.7, 38.8, 38.8, 38.8, 38.9, 39.1, 39.3, 39.3, 39.3, 39.5, 43.3, 43.3, 43.3, 43.5 ] }, { "x": 8.0, "output_throughput": { "median": 137.35, "min": 134.2, "max": 144.16, "n": 3 }, "median_ttft_ms": { "median": 126.25, "min": 111.65, "max": 127.12, "n": 3 }, "p99_ttft_ms": { "median": 678.42, "min": 640.97, "max": 687.86, "n": 3 }, "median_tpot_ms": { "median": 55.54, "min": 54.68, "max": 58.19, "n": 3 }, "p99_tpot_ms": { "median": 60.28, "min": 58.73, "max": 61.14, "n": 3 }, "p99_itl_ms": { "median": 61.68, "min": 60.21, "max": 65.62, "n": 3 }, "median_itl_ms": { "median": 54.65, "min": 54.49, "max": 57.64, "n": 3 }, "goodput_rps": { "median": 1.07, "min": 1.05, "max": 1.13, "n": 3 }, "request_throughput": { "median": 1.07, "min": 1.05, "max": 1.13, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 48.2, 49.7, 50.2, 52.2, 97.7, 97.8, 98.1, 99.4, 99.7, 99.9, 100.3, 100.4, 100.6, 101.0, 101.0, 101.1, 101.2, 101.3, 104.4, 104.5, 104.6, 104.6, 109.6, 109.9, 110.1, 110.4, 142.1, 153.0, 158.9, 374.1, 392.4, 392.6, 409.8, 410.4, 560.2, 560.3, 560.4, 574.6, 574.7, 574.8, 640.7, 640.7, 640.8, 640.9, 661.1, 661.2, 677.9, 678.3, 682.3, 682.8, 690.1 ], "tpot_quantiles_ms": [ 45.2, 45.2, 45.2, 47.3, 49.4, 53.1, 53.1, 53.8, 53.8, 53.9, 53.9, 53.9, 53.9, 54.6, 54.6, 54.6, 54.6, 55.4, 55.4, 55.5, 55.5, 55.5, 55.5, 56.1, 56.1, 56.1, 56.1, 56.1, 57.0, 57.0, 57.1, 57.1, 57.4, 57.8, 57.8, 57.8, 57.9, 58.2, 58.2, 58.2, 58.2, 58.2, 58.5, 58.5, 58.5, 58.7, 60.0, 60.0, 60.0, 60.0, 62.3 ] }, { "x": 16.0, "output_throughput": { "median": 189.07, "min": 188.77, "max": 193.57, "n": 3 }, "median_ttft_ms": { "median": 139.34, "min": 130.51, "max": 141.01, "n": 3 }, "p99_ttft_ms": { "median": 1228.67, "min": 1163.98, "max": 1452.08, "n": 3 }, "median_tpot_ms": { "median": 81.04, "min": 80.68, "max": 81.68, "n": 3 }, "p99_tpot_ms": { "median": 90.95, "min": 86.71, "max": 91.16, "n": 3 }, "p99_itl_ms": { "median": 94.48, "min": 93.34, "max": 97.03, "n": 3 }, "median_itl_ms": { "median": 79.16, "min": 78.31, "max": 79.44, "n": 3 }, "goodput_rps": { "median": 0.92, "min": 0.9, "max": 1.02, "n": 3 }, "request_throughput": { "median": 1.48, "min": 1.47, "max": 1.51, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 83.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 49.2, 51.1, 118.8, 119.0, 119.3, 119.4, 121.8, 122.1, 122.3, 123.4, 125.5, 125.8, 126.1, 127.0, 129.5, 129.9, 130.0, 130.2, 130.3, 130.5, 130.8, 131.3, 133.1, 133.6, 134.1, 135.0, 158.8, 373.9, 377.1, 382.4, 405.5, 632.4, 1081.8, 1082.1, 1082.4, 1082.7, 1105.7, 1106.9, 1108.0, 1120.4, 1121.2, 1122.3, 1163.5, 1163.8, 1163.9, 1228.4, 1228.6, 1450.0, 1450.8, 1451.8, 1452.2 ], "tpot_quantiles_ms": [ 71.2, 71.2, 71.2, 71.2, 71.8, 74.2, 74.3, 74.3, 74.3, 75.8, 75.8, 75.8, 75.9, 75.9, 75.9, 77.9, 77.9, 77.9, 79.5, 79.5, 79.5, 79.5, 80.7, 80.7, 80.7, 80.7, 81.1, 81.7, 81.7, 81.7, 81.7, 81.7, 82.0, 82.0, 82.0, 82.0, 83.0, 83.0, 83.0, 83.7, 84.9, 84.9, 84.9, 86.5, 86.5, 86.5, 86.5, 88.8, 89.6, 90.2, 93.3 ] }, { "x": 32.0, "output_throughput": { "median": 284.12, "min": 281.53, "max": 287.93, "n": 3 }, "median_ttft_ms": { "median": 176.21, "min": 169.98, "max": 253.55, "n": 3 }, "p99_ttft_ms": { "median": 2120.05, "min": 2106.71, "max": 3577.74, "n": 3 }, "median_tpot_ms": { "median": 104.92, "min": 102.37, "max": 105.59, "n": 3 }, "p99_tpot_ms": { "median": 120.63, "min": 119.26, "max": 123.88, "n": 3 }, "p99_itl_ms": { "median": 119.4, "min": 116.38, "max": 123.32, "n": 3 }, "median_itl_ms": { "median": 102.49, "min": 101.04, "max": 104.33, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.54, "n": 3 }, "request_throughput": { "median": 2.22, "min": 2.2, "max": 2.25, "n": 3 }, "completed": [ 128, 128, 128 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 52.7, 145.5, 148.6, 148.9, 149.3, 150.1, 150.8, 151.2, 152.2, 153.2, 154.2, 155.2, 155.8, 156.6, 158.5, 158.9, 159.6, 160.6, 169.5, 169.8, 169.9, 175.6, 175.9, 176.1, 188.8, 223.0, 325.6, 552.8, 579.7, 703.1, 1285.7, 1289.3, 1338.6, 1343.0, 1357.6, 1360.9, 1515.6, 1516.8, 1653.7, 1654.5, 2073.5, 2073.8, 2075.3, 2090.6, 2092.5, 2106.7, 2119.0, 2834.0, 2835.5, 3576.8, 3577.9 ], "tpot_quantiles_ms": [ 99.1, 99.1, 99.7, 99.7, 99.7, 99.7, 100.3, 100.3, 101.2, 101.2, 101.2, 101.3, 102.1, 102.1, 102.4, 102.4, 102.4, 102.4, 103.6, 103.6, 103.6, 103.6, 104.9, 104.9, 105.3, 105.3, 105.3, 105.4, 105.6, 106.0, 106.0, 106.1, 106.1, 107.0, 107.4, 107.9, 107.9, 108.2, 108.3, 108.3, 110.4, 110.4, 111.7, 111.7, 111.8, 111.8, 112.5, 118.0, 119.2, 121.1, 127.0 ] }, { "x": 64.0, "output_throughput": { "median": 442.71, "min": 440.94, "max": 447.09, "n": 3 }, "median_ttft_ms": { "median": 328.41, "min": 313.56, "max": 342.17, "n": 3 }, "p99_ttft_ms": { "median": 4010.74, "min": 3980.68, "max": 4029.3, "n": 3 }, "median_tpot_ms": { "median": 130.69, "min": 129.89, "max": 133.19, "n": 3 }, "p99_tpot_ms": { "median": 156.21, "min": 156.11, "max": 158.79, "n": 3 }, "p99_itl_ms": { "median": 151.21, "min": 150.16, "max": 156.14, "n": 3 }, "median_itl_ms": { "median": 127.26, "min": 126.1, "max": 127.51, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 3.46, "min": 3.44, "max": 3.49, "n": 3 }, "completed": [ 256, 256, 256 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.023 }, "ttft_quantiles_ms": [ 56.7, 138.4, 143.2, 155.5, 158.7, 167.8, 180.0, 180.3, 187.0, 197.3, 202.6, 218.2, 245.4, 256.0, 288.2, 293.6, 295.1, 295.6, 298.0, 300.1, 306.5, 309.9, 312.0, 327.8, 341.7, 342.1, 491.3, 1242.0, 1252.6, 1260.6, 1383.1, 1449.7, 1462.7, 1532.7, 2197.3, 2219.6, 2334.7, 2404.9, 2415.6, 2485.6, 3148.1, 3174.2, 3295.4, 3366.4, 3376.2, 3442.0, 3445.5, 3849.2, 3981.0, 4010.6, 4029.5 ], "tpot_quantiles_ms": [ 124.7, 125.8, 125.8, 126.1, 126.5, 126.6, 127.0, 127.1, 127.7, 127.8, 128.0, 128.4, 128.4, 128.4, 129.2, 129.3, 129.3, 129.6, 129.6, 129.7, 129.7, 130.2, 130.4, 130.7, 130.7, 130.7, 132.4, 132.7, 133.1, 133.1, 133.2, 133.4, 134.0, 134.4, 135.0, 136.1, 138.0, 140.0, 140.2, 141.0, 142.5, 143.7, 145.5, 147.5, 147.8, 148.5, 150.1, 150.1, 151.2, 156.1, 161.6 ] } ] }, "graphs-on-b": { "kv_pool_tokens": 233303, "points": [ { "x": 1.0, "output_throughput": { "median": 29.35, "min": 29.1, "max": 29.76, "n": 3 }, "median_ttft_ms": { "median": 270.12, "min": 269.81, "max": 307.12, "n": 3 }, "p99_ttft_ms": { "median": 298.41, "min": 297.32, "max": 424.18, "n": 3 }, "median_tpot_ms": { "median": 32.01, "min": 31.94, "max": 32.71, "n": 3 }, "p99_tpot_ms": { "median": 32.22, "min": 32.15, "max": 33.11, "n": 3 }, "p99_itl_ms": { "median": 34.83, "min": 34.18, "max": 42.36, "n": 3 }, "median_itl_ms": { "median": 31.93, "min": 31.92, "max": 32.56, "n": 3 }, "goodput_rps": { "median": 0.23, "min": 0.23, "max": 0.23, "n": 3 }, "request_throughput": { "median": 0.23, "min": 0.23, "max": 0.23, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.6, 41.6, 41.7, 66.2, 93.4, 209.1, 210.0, 248.3, 249.5, 251.3, 252.1, 255.1, 260.3, 261.0, 262.3, 268.3, 268.4, 268.4, 268.4, 269.2, 270.4, 271.8, 272.3, 276.3, 276.6, 276.7, 276.8, 277.0, 278.4, 282.6, 283.0, 284.4, 285.6, 286.5, 286.5, 288.1, 288.7, 298.7, 298.9, 300.2, 300.8, 301.7, 307.0, 307.2, 307.8, 312.2, 314.1, 338.6, 354.7, 369.7, 433.8 ], "tpot_quantiles_ms": [ 31.6, 31.6, 31.6, 31.7, 31.8, 31.8, 31.8, 31.8, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 31.9, 32.0, 32.0, 32.0, 32.0, 32.0, 32.0, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.1, 32.2, 32.2, 32.2, 32.2, 32.6, 32.7, 32.7, 32.7, 32.7, 32.7, 32.8, 32.8, 32.9, 32.9, 33.0, 33.1 ] }, { "x": 2.0, "output_throughput": { "median": 44.59, "min": 44.44, "max": 45.23, "n": 3 }, "median_ttft_ms": { "median": 72.21, "min": 68.79, "max": 73.83, "n": 3 }, "p99_ttft_ms": { "median": 82.13, "min": 76.07, "max": 85.58, "n": 3 }, "median_tpot_ms": { "median": 44.61, "min": 44.54, "max": 45.22, "n": 3 }, "p99_tpot_ms": { "median": 45.98, "min": 45.93, "max": 46.15, "n": 3 }, "p99_itl_ms": { "median": 49.09, "min": 48.46, "max": 55.67, "n": 3 }, "median_itl_ms": { "median": 44.83, "min": 44.42, "max": 45.07, "n": 3 }, "goodput_rps": { "median": 0.35, "min": 0.35, "max": 0.35, "n": 3 }, "request_throughput": { "median": 0.35, "min": 0.35, "max": 0.35, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 60.1, 60.1, 60.5, 60.8, 62.5, 62.6, 63.0, 63.3, 63.7, 63.9, 63.9, 64.0, 64.1, 64.2, 64.5, 64.6, 64.8, 64.8, 64.9, 65.0, 65.1, 67.0, 71.0, 71.3, 71.7, 72.7, 72.7, 72.8, 73.1, 73.1, 73.2, 73.4, 73.4, 73.7, 73.7, 73.7, 74.4, 74.8, 74.8, 75.0, 75.2, 75.3, 76.0, 76.3, 76.4, 76.5, 76.9, 77.3, 83.0, 83.0, 86.0 ], "tpot_quantiles_ms": [ 40.8, 40.8, 40.9, 41.7, 41.8, 42.5, 42.5, 43.8, 43.9, 43.9, 43.9, 44.0, 44.0, 44.0, 44.1, 44.1, 44.2, 44.2, 44.3, 44.4, 44.4, 44.5, 44.5, 44.6, 44.6, 44.7, 44.7, 44.7, 44.8, 44.8, 44.9, 44.9, 45.1, 45.2, 45.2, 45.3, 45.3, 45.3, 45.3, 45.4, 45.4, 45.4, 45.4, 45.4, 45.4, 45.9, 45.9, 45.9, 46.0, 46.1, 46.2 ] }, { "x": 4.0, "output_throughput": { "median": 57.69, "min": 57.55, "max": 58.06, "n": 3 }, "median_ttft_ms": { "median": 114.75, "min": 110.29, "max": 115.33, "n": 3 }, "p99_ttft_ms": { "median": 116.23, "min": 114.25, "max": 116.8, "n": 3 }, "median_tpot_ms": { "median": 69.5, "min": 69.31, "max": 69.73, "n": 3 }, "p99_tpot_ms": { "median": 72.27, "min": 70.44, "max": 73.45, "n": 3 }, "p99_itl_ms": { "median": 78.62, "min": 75.37, "max": 81.87, "n": 3 }, "median_itl_ms": { "median": 68.97, "min": 68.93, "max": 69.25, "n": 3 }, "goodput_rps": { "median": 0.45, "min": 0.45, "max": 0.45, "n": 3 }, "request_throughput": { "median": 0.45, "min": 0.45, "max": 0.45, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 46.6, 46.6, 47.0, 47.8, 47.9, 48.1, 48.1, 48.5, 48.8, 49.2, 50.9, 50.9, 51.0, 108.0, 108.1, 108.2, 110.3, 110.3, 110.3, 110.4, 111.0, 111.0, 111.2, 112.4, 112.5, 112.5, 114.1, 114.2, 114.3, 114.3, 114.6, 114.6, 114.7, 114.9, 114.9, 115.1, 115.2, 115.5, 115.7, 115.7, 115.7, 115.7, 115.7, 115.9, 116.0, 116.0, 116.1, 116.2, 116.6, 116.7, 116.8 ], "tpot_quantiles_ms": [ 63.8, 63.8, 63.8, 63.8, 64.3, 65.2, 65.2, 65.2, 65.7, 66.6, 66.6, 66.6, 67.2, 67.9, 67.9, 67.9, 68.4, 68.4, 69.3, 69.3, 69.3, 69.3, 69.3, 69.3, 69.5, 69.5, 69.5, 69.7, 69.7, 69.7, 69.8, 69.8, 70.0, 70.0, 70.0, 70.0, 70.0, 70.2, 70.2, 70.2, 70.3, 70.5, 70.8, 71.9, 71.9, 71.9, 72.3, 73.0, 73.0, 73.0, 73.5 ] }, { "x": 8.0, "output_throughput": { "median": 75.52, "min": 72.28, "max": 78.68, "n": 3 }, "median_ttft_ms": { "median": 213.96, "min": 184.71, "max": 221.37, "n": 3 }, "p99_ttft_ms": { "median": 1086.01, "min": 970.22, "max": 1189.48, "n": 3 }, "median_tpot_ms": { "median": 101.26, "min": 99.28, "max": 108.89, "n": 3 }, "p99_tpot_ms": { "median": 112.83, "min": 106.98, "max": 113.32, "n": 3 }, "p99_itl_ms": { "median": 116.8, "min": 116.77, "max": 121.41, "n": 3 }, "median_itl_ms": { "median": 100.76, "min": 99.79, "max": 107.09, "n": 3 }, "goodput_rps": { "median": 0.17, "min": 0.0, "max": 0.41, "n": 3 }, "request_throughput": { "median": 0.59, "min": 0.56, "max": 0.61, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 49.0, 49.4, 51.4, 64.5, 146.6, 147.0, 147.3, 148.0, 148.0, 148.1, 148.3, 148.9, 149.3, 149.4, 149.7, 158.9, 159.0, 159.1, 162.9, 163.1, 163.4, 163.6, 166.4, 166.5, 166.5, 166.6, 264.3, 285.7, 319.8, 615.9, 616.3, 949.5, 949.6, 949.7, 949.7, 969.5, 969.9, 970.1, 991.9, 992.3, 993.0, 993.2, 1032.4, 1032.8, 1033.2, 1033.4, 1085.9, 1086.0, 1189.0, 1189.2, 1189.5 ], "tpot_quantiles_ms": [ 82.3, 82.3, 82.3, 82.3, 89.4, 98.9, 98.9, 98.9, 98.9, 98.9, 98.9, 98.9, 99.3, 99.3, 99.3, 99.3, 99.6, 99.6, 99.6, 99.7, 102.2, 102.2, 102.2, 102.2, 103.4, 103.4, 104.3, 105.6, 105.6, 105.6, 105.6, 106.1, 106.1, 106.1, 106.1, 106.9, 107.1, 107.1, 108.9, 108.9, 108.9, 108.9, 109.0, 109.0, 109.0, 109.7, 112.2, 112.2, 112.2, 112.2, 114.7 ] }, { "x": 16.0, "output_throughput": { "median": 103.44, "min": 102.41, "max": 106.7, "n": 3 }, "median_ttft_ms": { "median": 257.58, "min": 252.87, "max": 276.73, "n": 3 }, "p99_ttft_ms": { "median": 1693.72, "min": 1643.27, "max": 1731.25, "n": 3 }, "median_tpot_ms": { "median": 152.17, "min": 149.59, "max": 154.34, "n": 3 }, "p99_tpot_ms": { "median": 158.01, "min": 155.49, "max": 161.61, "n": 3 }, "p99_itl_ms": { "median": 164.39, "min": 162.51, "max": 169.18, "n": 3 }, "median_itl_ms": { "median": 149.97, "min": 145.07, "max": 149.98, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.81, "min": 0.8, "max": 0.83, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.011 }, "ttft_quantiles_ms": [ 50.2, 53.0, 184.5, 184.7, 184.9, 184.9, 185.0, 185.1, 185.3, 196.8, 197.6, 197.8, 197.9, 198.1, 198.3, 199.0, 202.4, 202.5, 202.6, 202.7, 234.1, 234.4, 234.8, 235.4, 246.3, 249.9, 296.3, 619.1, 667.8, 678.8, 1461.4, 1462.2, 1462.9, 1463.6, 1642.6, 1642.8, 1643.1, 1643.3, 1670.4, 1671.3, 1672.7, 1685.4, 1686.4, 1687.4, 1691.3, 1691.6, 1692.7, 1693.8, 1729.4, 1730.2, 1731.3 ], "tpot_quantiles_ms": [ 132.9, 132.9, 132.9, 132.9, 134.0, 139.8, 139.8, 139.9, 139.9, 142.0, 142.0, 142.0, 142.1, 142.1, 142.1, 142.1, 145.8, 145.8, 145.8, 148.9, 149.0, 149.0, 149.0, 149.6, 149.6, 149.6, 149.6, 151.9, 151.9, 152.2, 152.2, 152.2, 152.4, 152.4, 152.4, 152.5, 152.5, 152.5, 152.8, 154.2, 154.3, 154.3, 154.4, 154.4, 157.2, 157.4, 157.6, 157.6, 157.6, 160.2, 163.6 ] }, { "x": 32.0, "output_throughput": { "median": 154.64, "min": 152.37, "max": 155.89, "n": 3 }, "median_ttft_ms": { "median": 400.92, "min": 350.71, "max": 430.56, "n": 3 }, "p99_ttft_ms": { "median": 3071.46, "min": 3030.63, "max": 3133.06, "n": 3 }, "median_tpot_ms": { "median": 195.94, "min": 194.32, "max": 200.27, "n": 3 }, "p99_tpot_ms": { "median": 219.65, "min": 209.43, "max": 224.67, "n": 3 }, "p99_itl_ms": { "median": 219.54, "min": 215.91, "max": 219.72, "n": 3 }, "median_itl_ms": { "median": 194.79, "min": 191.95, "max": 198.4, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.21, "min": 1.19, "max": 1.22, "n": 3 }, "completed": [ 128, 128, 128 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 88.0, "thermal_share": 0.042 }, "ttft_quantiles_ms": [ 53.3, 173.9, 177.5, 201.2, 228.9, 229.4, 230.6, 240.4, 240.6, 240.7, 241.6, 242.7, 250.4, 262.1, 265.6, 265.7, 265.7, 275.4, 324.5, 337.4, 337.5, 350.7, 371.5, 391.0, 400.0, 430.4, 656.0, 691.0, 1935.8, 1939.8, 1945.6, 1946.9, 1977.7, 1980.6, 1982.0, 1983.0, 1984.5, 1988.8, 1992.1, 2042.3, 2043.5, 2976.9, 3015.0, 3017.3, 3026.4, 3027.7, 3029.2, 3071.1, 3071.4, 3132.0, 3133.1 ], "tpot_quantiles_ms": [ 185.7, 185.7, 188.4, 188.4, 190.6, 190.6, 191.7, 191.7, 192.6, 193.1, 193.1, 193.4, 193.4, 193.4, 193.5, 194.1, 194.1, 194.3, 194.3, 194.8, 194.8, 194.8, 194.8, 195.2, 195.9, 196.0, 196.9, 196.9, 199.4, 200.3, 200.3, 200.3, 201.2, 201.2, 201.4, 201.4, 201.4, 201.4, 206.3, 206.3, 206.8, 207.6, 207.6, 208.7, 210.4, 210.4, 210.4, 214.4, 214.4, 214.5, 228.1 ] }, { "x": 64.0, "output_throughput": { "median": 245.09, "min": 239.36, "max": 247.7, "n": 3 }, "median_ttft_ms": { "median": 572.62, "min": 541.65, "max": 589.76, "n": 3 }, "p99_ttft_ms": { "median": 5560.9, "min": 5379.53, "max": 5750.67, "n": 3 }, "median_tpot_ms": { "median": 245.98, "min": 241.46, "max": 246.56, "n": 3 }, "p99_tpot_ms": { "median": 272.95, "min": 271.92, "max": 291.03, "n": 3 }, "p99_itl_ms": { "median": 270.76, "min": 267.51, "max": 273.3, "n": 3 }, "median_itl_ms": { "median": 239.3, "min": 237.22, "max": 244.16, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.91, "min": 1.87, "max": 1.94, "n": 3 }, "completed": [ 256, 256, 256 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 88.0, "thermal_share": 0.074 }, "ttft_quantiles_ms": [ 59.7, 193.6, 231.1, 232.2, 235.3, 274.7, 276.7, 277.1, 283.3, 294.2, 296.1, 371.7, 404.4, 417.1, 417.5, 429.3, 429.5, 501.8, 504.8, 508.0, 524.0, 525.0, 541.6, 556.6, 572.6, 589.6, 1562.3, 1599.2, 1611.3, 1614.5, 1843.6, 1943.8, 1953.7, 2833.5, 2881.1, 2889.5, 3164.7, 3198.5, 3266.9, 4103.7, 4144.2, 4166.8, 4487.8, 4497.7, 4599.4, 5149.6, 5330.4, 5379.0, 5427.9, 5561.1, 5751.1 ], "tpot_quantiles_ms": [ 233.0, 235.2, 235.2, 236.4, 236.5, 237.1, 237.5, 239.0, 239.0, 239.1, 239.1, 240.2, 240.2, 240.3, 241.4, 241.4, 241.5, 241.5, 241.6, 242.4, 242.8, 242.8, 243.5, 243.6, 244.6, 245.0, 245.6, 246.0, 246.0, 246.2, 246.6, 247.6, 247.6, 251.6, 251.8, 253.3, 253.4, 255.0, 256.3, 259.8, 261.5, 261.9, 263.9, 265.2, 266.3, 270.3, 271.9, 271.9, 274.4, 280.7, 294.2 ] } ] }, "graphs-off-b": { "kv_pool_tokens": 234866, "points": [ { "x": 1.0, "output_throughput": { "median": 29.17, "min": 28.71, "max": 29.4, "n": 3 }, "median_ttft_ms": { "median": 289.79, "min": 287.99, "max": 300.14, "n": 3 }, "p99_ttft_ms": { "median": 319.72, "min": 316.93, "max": 332.17, "n": 3 }, "median_tpot_ms": { "median": 32.27, "min": 32.27, "max": 32.96, "n": 3 }, "p99_tpot_ms": { "median": 32.36, "min": 32.34, "max": 33.06, "n": 3 }, "p99_itl_ms": { "median": 35.1, "min": 35.01, "max": 35.58, "n": 3 }, "median_itl_ms": { "median": 32.31, "min": 32.29, "max": 32.98, "n": 3 }, "goodput_rps": { "median": 0.23, "min": 0.22, "max": 0.23, "n": 3 }, "request_throughput": { "median": 0.23, "min": 0.22, "max": 0.23, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 59.7, 59.7, 61.5, 69.5, 95.0, 222.1, 226.7, 263.2, 268.0, 268.6, 269.3, 276.2, 277.3, 278.1, 279.4, 283.6, 285.7, 285.7, 285.9, 289.1, 289.2, 289.5, 290.0, 290.4, 292.5, 292.6, 293.5, 293.7, 294.4, 295.5, 296.3, 296.3, 297.8, 298.0, 298.0, 300.1, 300.2, 301.0, 303.8, 305.2, 305.6, 306.9, 307.2, 307.3, 308.0, 311.7, 318.6, 321.0, 321.8, 328.2, 332.9 ], "tpot_quantiles_ms": [ 32.0, 32.0, 32.1, 32.2, 32.2, 32.2, 32.2, 32.2, 32.2, 32.2, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.3, 32.4, 32.4, 32.4, 32.4, 32.6, 32.9, 32.9, 32.9, 32.9, 32.9, 33.0, 33.0, 33.0, 33.0, 33.0, 33.0, 33.0, 33.0, 33.1 ] }, { "x": 2.0, "output_throughput": { "median": 43.66, "min": 43.38, "max": 44.31, "n": 3 }, "median_ttft_ms": { "median": 78.75, "min": 78.45, "max": 79.66, "n": 3 }, "p99_ttft_ms": { "median": 98.66, "min": 96.34, "max": 104.68, "n": 3 }, "median_tpot_ms": { "median": 45.49, "min": 45.45, "max": 45.95, "n": 3 }, "p99_tpot_ms": { "median": 46.81, "min": 46.53, "max": 47.22, "n": 3 }, "p99_itl_ms": { "median": 49.79, "min": 49.71, "max": 51.18, "n": 3 }, "median_itl_ms": { "median": 45.64, "min": 45.34, "max": 46.06, "n": 3 }, "goodput_rps": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "request_throughput": { "median": 0.34, "min": 0.34, "max": 0.35, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 60.4, 60.4, 62.6, 62.8, 62.8, 62.9, 63.1, 63.3, 63.4, 63.7, 63.7, 63.8, 64.0, 64.1, 64.1, 64.2, 64.2, 64.2, 64.4, 64.4, 64.5, 64.5, 64.6, 65.3, 65.7, 66.3, 91.6, 93.0, 93.0, 93.4, 93.7, 94.2, 94.4, 94.4, 94.4, 94.5, 94.5, 94.6, 94.9, 94.9, 95.2, 95.5, 95.6, 95.7, 96.4, 96.5, 96.5, 96.8, 98.2, 98.7, 106.4 ], "tpot_quantiles_ms": [ 41.7, 41.7, 42.0, 42.3, 42.6, 43.1, 43.4, 44.5, 44.6, 44.7, 44.8, 44.9, 45.0, 45.1, 45.1, 45.2, 45.2, 45.2, 45.4, 45.4, 45.4, 45.5, 45.5, 45.6, 45.6, 45.6, 45.6, 45.7, 45.8, 45.8, 45.8, 45.8, 45.8, 45.9, 45.9, 46.0, 46.1, 46.2, 46.2, 46.3, 46.3, 46.4, 46.5, 46.5, 46.5, 46.6, 46.8, 46.8, 47.0, 47.1, 47.2 ] }, { "x": 4.0, "output_throughput": { "median": 56.83, "min": 56.54, "max": 57.2, "n": 3 }, "median_ttft_ms": { "median": 122.02, "min": 121.7, "max": 123.39, "n": 3 }, "p99_ttft_ms": { "median": 124.23, "min": 123.84, "max": 146.57, "n": 3 }, "median_tpot_ms": { "median": 70.51, "min": 70.35, "max": 70.58, "n": 3 }, "p99_tpot_ms": { "median": 73.29, "min": 71.14, "max": 74.53, "n": 3 }, "p99_itl_ms": { "median": 77.29, "min": 75.11, "max": 77.69, "n": 3 }, "median_itl_ms": { "median": 70.27, "min": 70.09, "max": 70.49, "n": 3 }, "goodput_rps": { "median": 0.44, "min": 0.44, "max": 0.45, "n": 3 }, "request_throughput": { "median": 0.44, "min": 0.44, "max": 0.45, "n": 3 }, "completed": [ 16, 16, 16 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 62.9, 62.9, 64.0, 64.5, 65.6, 65.6, 66.5, 66.7, 66.8, 67.1, 67.8, 68.2, 81.5, 119.5, 119.5, 119.6, 120.9, 120.9, 121.0, 121.1, 121.1, 121.2, 121.5, 121.6, 121.8, 121.9, 121.9, 122.1, 122.2, 122.4, 122.5, 122.5, 122.7, 122.7, 122.7, 122.9, 123.3, 123.5, 123.6, 123.7, 123.7, 123.8, 123.8, 123.8, 123.8, 124.1, 124.2, 124.2, 146.2, 146.5, 146.6 ], "tpot_quantiles_ms": [ 64.9, 64.9, 64.9, 64.9, 65.3, 66.4, 66.4, 66.4, 66.8, 67.8, 67.8, 67.8, 68.2, 69.1, 69.1, 69.1, 69.5, 69.5, 70.0, 70.0, 70.0, 70.3, 70.3, 70.3, 70.4, 70.4, 70.4, 70.4, 70.7, 70.7, 70.7, 70.7, 70.7, 70.7, 70.7, 70.8, 70.9, 71.2, 71.2, 71.5, 71.5, 71.5, 71.9, 72.9, 72.9, 72.9, 73.4, 74.1, 74.1, 74.1, 74.6 ] }, { "x": 8.0, "output_throughput": { "median": 74.7, "min": 71.73, "max": 77.9, "n": 3 }, "median_ttft_ms": { "median": 231.66, "min": 201.39, "max": 255.89, "n": 3 }, "p99_ttft_ms": { "median": 1005.21, "min": 992.39, "max": 1041.44, "n": 3 }, "median_tpot_ms": { "median": 102.67, "min": 100.89, "max": 109.95, "n": 3 }, "p99_tpot_ms": { "median": 113.79, "min": 108.86, "max": 114.69, "n": 3 }, "p99_itl_ms": { "median": 118.14, "min": 113.28, "max": 119.09, "n": 3 }, "median_itl_ms": { "median": 102.18, "min": 102.03, "max": 108.27, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.0, "max": 0.13, "n": 3 }, "request_throughput": { "median": 0.58, "min": 0.56, "max": 0.61, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 64.7, 66.4, 67.6, 87.9, 159.3, 159.6, 159.7, 161.4, 161.5, 161.6, 161.6, 161.7, 161.7, 162.0, 162.3, 173.5, 173.6, 173.6, 173.7, 173.8, 173.9, 173.9, 203.3, 203.4, 203.5, 203.5, 289.3, 310.3, 327.9, 893.5, 893.9, 894.1, 894.2, 958.3, 958.4, 958.5, 980.7, 980.9, 981.2, 981.4, 992.2, 992.3, 992.3, 992.4, 1004.5, 1004.9, 1005.1, 1040.6, 1041.1, 1041.3, 1041.5 ], "tpot_quantiles_ms": [ 83.3, 83.3, 83.3, 83.3, 88.6, 99.4, 99.4, 99.4, 100.1, 100.1, 100.1, 100.1, 100.7, 100.7, 100.7, 100.7, 100.9, 100.9, 100.9, 100.9, 103.7, 103.7, 103.7, 103.7, 104.1, 104.1, 104.1, 104.1, 106.2, 106.4, 106.4, 106.4, 106.4, 107.7, 107.7, 107.7, 107.7, 109.0, 110.5, 110.5, 110.5, 110.5, 110.6, 110.6, 110.6, 110.6, 113.2, 113.2, 113.2, 113.2, 116.2 ] }, { "x": 16.0, "output_throughput": { "median": 102.62, "min": 101.09, "max": 104.54, "n": 3 }, "median_ttft_ms": { "median": 264.1, "min": 244.25, "max": 276.57, "n": 3 }, "p99_ttft_ms": { "median": 1717.63, "min": 1645.32, "max": 1748.84, "n": 3 }, "median_tpot_ms": { "median": 153.03, "min": 152.37, "max": 156.1, "n": 3 }, "p99_tpot_ms": { "median": 159.97, "min": 158.52, "max": 161.7, "n": 3 }, "p99_itl_ms": { "median": 165.41, "min": 163.62, "max": 169.0, "n": 3 }, "median_itl_ms": { "median": 151.61, "min": 148.2, "max": 151.82, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.8, "min": 0.79, "max": 0.82, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 68.3, 71.1, 176.9, 177.2, 177.9, 189.0, 189.1, 190.1, 208.3, 209.4, 210.7, 222.3, 222.4, 222.5, 222.7, 224.2, 224.4, 224.5, 224.5, 239.7, 239.9, 240.0, 240.2, 243.4, 244.2, 262.7, 303.5, 1212.0, 1212.2, 1212.6, 1248.8, 1249.3, 1249.7, 1258.9, 1259.5, 1259.8, 1301.2, 1301.7, 1302.1, 1482.6, 1482.8, 1483.0, 1483.2, 1583.9, 1584.1, 1584.2, 1584.4, 1645.3, 1685.3, 1717.7, 1748.9 ], "tpot_quantiles_ms": [ 135.2, 135.6, 135.6, 135.6, 136.6, 141.2, 141.8, 141.8, 141.8, 143.9, 145.0, 147.7, 147.7, 147.7, 148.1, 148.1, 148.1, 148.1, 150.7, 150.7, 150.7, 151.3, 151.3, 151.3, 151.3, 152.4, 152.4, 152.4, 152.4, 153.0, 154.3, 154.3, 154.3, 154.7, 155.1, 155.1, 155.1, 155.1, 156.2, 156.2, 156.2, 156.2, 156.9, 156.9, 156.9, 157.4, 159.3, 159.6, 159.6, 159.6, 164.4 ] }, { "x": 32.0, "output_throughput": { "median": 153.32, "min": 152.7, "max": 154.59, "n": 3 }, "median_ttft_ms": { "median": 380.49, "min": 323.27, "max": 393.22, "n": 3 }, "p99_ttft_ms": { "median": 3071.63, "min": 2822.53, "max": 3077.48, "n": 3 }, "median_tpot_ms": { "median": 196.71, "min": 195.45, "max": 201.85, "n": 3 }, "p99_tpot_ms": { "median": 214.17, "min": 209.92, "max": 217.11, "n": 3 }, "p99_itl_ms": { "median": 213.34, "min": 211.81, "max": 218.43, "n": 3 }, "median_itl_ms": { "median": 197.55, "min": 193.85, "max": 199.5, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.2, "min": 1.19, "max": 1.21, "n": 3 }, "completed": [ 128, 128, 128 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 89.0, "thermal_share": 0.025 }, "ttft_quantiles_ms": [ 72.9, 224.4, 225.4, 250.1, 251.4, 252.8, 253.8, 254.1, 254.2, 254.4, 255.7, 256.8, 258.5, 260.2, 270.5, 281.3, 303.4, 303.4, 303.9, 322.6, 362.3, 363.2, 370.2, 379.5, 380.3, 393.1, 1212.0, 1215.0, 1218.0, 1397.4, 1400.0, 1572.5, 1573.6, 1608.1, 1610.1, 1640.5, 1641.6, 2498.5, 2501.6, 2526.1, 2528.9, 2671.6, 2672.7, 2727.0, 2727.3, 2820.9, 2822.3, 2875.9, 2876.3, 3040.1, 3077.6 ], "tpot_quantiles_ms": [ 187.9, 190.0, 192.2, 192.2, 192.4, 192.5, 192.9, 192.9, 193.5, 193.5, 194.3, 194.3, 194.9, 195.4, 195.4, 195.5, 196.1, 196.1, 196.1, 196.2, 196.2, 196.7, 196.7, 196.7, 196.7, 197.3, 198.4, 201.0, 201.8, 201.9, 201.9, 202.4, 202.4, 202.6, 203.8, 204.4, 205.1, 205.1, 205.9, 205.9, 207.0, 207.0, 208.4, 208.4, 208.4, 208.4, 214.1, 214.2, 214.2, 217.1, 225.4 ] }, { "x": 64.0, "output_throughput": { "median": 242.78, "min": 239.93, "max": 245.54, "n": 3 }, "median_ttft_ms": { "median": 549.62, "min": 523.86, "max": 550.6, "n": 3 }, "p99_ttft_ms": { "median": 5533.14, "min": 5519.44, "max": 5669.59, "n": 3 }, "median_tpot_ms": { "median": 247.44, "min": 242.0, "max": 247.46, "n": 3 }, "p99_tpot_ms": { "median": 274.32, "min": 274.28, "max": 280.35, "n": 3 }, "p99_itl_ms": { "median": 253.97, "min": 249.57, "max": 258.59, "n": 3 }, "median_itl_ms": { "median": 241.9, "min": 238.91, "max": 244.39, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.9, "min": 1.87, "max": 1.92, "n": 3 }, "completed": [ 256, 256, 256 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.021 }, "ttft_quantiles_ms": [ 84.0, 193.8, 207.6, 289.8, 290.3, 299.7, 336.6, 354.0, 360.2, 367.9, 373.7, 376.5, 379.7, 385.2, 480.0, 480.2, 494.2, 494.3, 513.9, 521.0, 540.1, 542.0, 542.3, 542.5, 545.7, 549.7, 1083.1, 1320.2, 1376.1, 1554.3, 1641.2, 2308.8, 2389.9, 2618.0, 2668.3, 2853.8, 2927.8, 3610.4, 3686.2, 3906.4, 3981.5, 4162.7, 4230.1, 4924.3, 4988.0, 5223.3, 5231.8, 5278.3, 5423.0, 5460.8, 5669.9 ], "tpot_quantiles_ms": [ 234.9, 236.6, 237.6, 237.8, 237.8, 239.1, 240.2, 240.2, 240.3, 240.6, 240.6, 241.8, 241.9, 241.9, 242.0, 242.0, 242.0, 242.1, 243.3, 243.5, 243.5, 244.4, 244.9, 245.1, 246.5, 246.5, 247.4, 247.4, 247.4, 247.5, 247.8, 249.0, 249.0, 251.5, 252.4, 253.7, 253.8, 257.8, 258.1, 259.8, 262.5, 264.0, 264.1, 268.0, 268.3, 270.1, 272.8, 274.3, 274.3, 280.3, 290.5 ] } ] } }, "prefix": { "default": { "kv_pool_tokens": 240466, "points": [ { "x": 0.0, "output_throughput": { "median": 121.01, "min": 116.52, "max": 124.85, "n": 3 }, "median_ttft_ms": { "median": 487.18, "min": 485.8, "max": 491.36, "n": 3 }, "p99_ttft_ms": { "median": 738.64, "min": 732.7, "max": 765.06, "n": 3 }, "median_tpot_ms": { "median": 58.09, "min": 56.8, "max": 61.47, "n": 3 }, "p99_tpot_ms": { "median": 72.32, "min": 70.48, "max": 76.28, "n": 3 }, "p99_itl_ms": { "median": 72.18, "min": 71.91, "max": 75.71, "n": 3 }, "median_itl_ms": { "median": 56.4, "min": 54.54, "max": 59.69, "n": 3 }, "goodput_rps": { "median": 1.89, "min": 1.82, "max": 1.95, "n": 3 }, "request_throughput": { "median": 1.89, "min": 1.82, "max": 1.95, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 48.2, 202.5, 208.6, 211.5, 213.2, 214.3, 218.8, 369.2, 384.0, 384.2, 398.2, 398.5, 455.0, 456.7, 462.3, 465.9, 467.8, 468.6, 479.4, 482.6, 482.9, 483.3, 483.5, 484.0, 488.3, 488.5, 490.0, 490.6, 491.3, 524.1, 524.5, 527.8, 528.1, 532.9, 533.0, 723.2, 725.1, 726.0, 727.0, 727.1, 727.2, 729.5, 732.6, 732.7, 733.2, 733.5, 734.5, 735.0, 737.0, 738.9, 807.1 ], "tpot_quantiles_ms": [ 41.6, 41.7, 45.4, 45.9, 47.8, 48.1, 48.9, 49.7, 50.5, 52.0, 52.4, 53.0, 53.0, 53.1, 54.3, 54.4, 55.2, 55.8, 56.3, 56.9, 56.9, 57.6, 58.0, 58.0, 58.1, 59.5, 59.7, 60.0, 60.2, 60.7, 60.7, 60.9, 61.7, 63.3, 63.3, 63.5, 63.9, 64.1, 65.0, 65.2, 67.6, 68.4, 68.5, 68.5, 68.5, 69.1, 70.9, 70.9, 71.7, 71.7, 76.8 ] }, { "x": 256.0, "output_throughput": { "median": 127.71, "min": 123.85, "max": 147.2, "n": 3 }, "median_ttft_ms": { "median": 583.39, "min": 512.26, "max": 677.6, "n": 3 }, "p99_ttft_ms": { "median": 695.14, "min": 676.04, "max": 713.91, "n": 3 }, "median_tpot_ms": { "median": 50.39, "min": 49.84, "max": 59.1, "n": 3 }, "p99_tpot_ms": { "median": 70.03, "min": 60.02, "max": 80.89, "n": 3 }, "p99_itl_ms": { "median": 71.19, "min": 65.37, "max": 81.56, "n": 3 }, "median_itl_ms": { "median": 47.85, "min": 46.53, "max": 56.34, "n": 3 }, "goodput_rps": { "median": 2.0, "min": 1.94, "max": 2.3, "n": 3 }, "request_throughput": { "median": 2.0, "min": 1.94, "max": 2.3, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 51.2, 185.5, 200.5, 201.1, 203.3, 206.6, 208.0, 411.2, 414.9, 422.1, 425.3, 436.5, 448.6, 454.6, 458.4, 460.3, 495.1, 495.3, 503.7, 504.0, 509.0, 512.1, 512.3, 512.4, 512.7, 516.8, 517.1, 517.2, 517.6, 583.3, 583.6, 626.0, 672.2, 675.9, 676.0, 677.5, 678.7, 680.8, 682.1, 683.0, 688.8, 689.5, 690.0, 690.9, 691.1, 692.0, 692.6, 693.0, 693.7, 694.4, 747.0 ], "tpot_quantiles_ms": [ 39.6, 39.6, 39.8, 39.8, 41.0, 41.5, 41.6, 41.6, 42.4, 42.4, 44.5, 44.8, 44.8, 45.5, 45.6, 46.7, 48.0, 48.0, 49.1, 49.1, 49.8, 49.9, 49.9, 50.2, 50.2, 50.4, 50.5, 51.0, 51.0, 53.5, 55.3, 55.5, 55.5, 57.0, 57.0, 57.7, 57.7, 59.5, 60.7, 61.2, 62.1, 62.1, 63.1, 63.3, 66.2, 66.2, 66.7, 66.7, 75.3, 75.4, 83.1 ] }, { "x": 1024.0, "output_throughput": { "median": 136.83, "min": 135.2, "max": 137.28, "n": 3 }, "median_ttft_ms": { "median": 545.85, "min": 545.84, "max": 702.81, "n": 3 }, "p99_ttft_ms": { "median": 1181.72, "min": 1134.73, "max": 1407.17, "n": 3 }, "median_tpot_ms": { "median": 50.95, "min": 47.04, "max": 51.02, "n": 3 }, "p99_tpot_ms": { "median": 65.55, "min": 62.08, "max": 66.73, "n": 3 }, "p99_itl_ms": { "median": 67.93, "min": 64.01, "max": 72.41, "n": 3 }, "median_itl_ms": { "median": 48.55, "min": 44.8, "max": 49.57, "n": 3 }, "goodput_rps": { "median": 1.88, "min": 1.77, "max": 1.91, "n": 3 }, "request_throughput": { "median": 2.14, "min": 2.11, "max": 2.15, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 54.8, 200.3, 203.8, 211.7, 241.8, 244.3, 250.8, 402.4, 447.4, 447.8, 454.1, 539.0, 539.9, 540.3, 540.7, 541.1, 542.4, 542.6, 542.7, 542.9, 543.2, 545.4, 545.8, 545.9, 546.5, 546.7, 546.9, 547.3, 547.7, 549.8, 550.7, 554.4, 556.6, 557.1, 599.1, 692.3, 692.7, 701.4, 702.7, 703.0, 703.3, 703.9, 716.2, 779.8, 1107.6, 1107.8, 1181.5, 1181.7, 1210.7, 1407.0, 1407.2 ], "tpot_quantiles_ms": [ 39.9, 40.1, 41.9, 42.0, 42.1, 42.9, 42.9, 44.2, 44.5, 44.5, 44.5, 44.5, 44.9, 46.0, 46.0, 46.8, 46.8, 47.0, 47.0, 48.1, 48.1, 48.2, 48.9, 48.9, 48.9, 48.9, 49.2, 49.4, 49.9, 50.9, 51.0, 51.0, 51.2, 51.2, 51.5, 51.5, 52.8, 53.1, 53.6, 53.6, 53.8, 53.8, 54.4, 54.4, 55.0, 56.7, 57.2, 60.9, 60.9, 64.2, 68.8 ] }, { "x": 2048.0, "output_throughput": { "median": 138.65, "min": 134.06, "max": 141.54, "n": 3 }, "median_ttft_ms": { "median": 564.06, "min": 561.74, "max": 570.93, "n": 3 }, "p99_ttft_ms": { "median": 1879.27, "min": 1590.77, "max": 2065.63, "n": 3 }, "median_tpot_ms": { "median": 47.97, "min": 47.85, "max": 50.9, "n": 3 }, "p99_tpot_ms": { "median": 64.88, "min": 61.57, "max": 67.24, "n": 3 }, "p99_itl_ms": { "median": 69.73, "min": 68.54, "max": 71.38, "n": 3 }, "median_itl_ms": { "median": 44.61, "min": 43.66, "max": 48.15, "n": 3 }, "goodput_rps": { "median": 1.93, "min": 1.87, "max": 2.07, "n": 3 }, "request_throughput": { "median": 2.17, "min": 2.09, "max": 2.21, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 83.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 57.0, 199.7, 214.2, 242.4, 243.3, 251.7, 259.9, 552.4, 553.9, 554.3, 555.1, 556.6, 558.1, 559.8, 560.5, 560.7, 561.0, 561.1, 561.3, 561.4, 561.8, 562.3, 562.7, 563.1, 563.2, 563.6, 563.9, 564.1, 564.4, 566.4, 567.2, 567.5, 567.8, 568.6, 569.7, 571.3, 572.6, 573.2, 573.5, 573.8, 575.6, 576.5, 698.6, 698.9, 717.6, 810.3, 1133.0, 1137.0, 1369.1, 1879.3, 2065.6 ], "tpot_quantiles_ms": [ 40.6, 40.6, 42.2, 42.3, 42.4, 42.4, 42.4, 42.5, 43.3, 43.3, 43.7, 43.8, 43.8, 43.8, 43.9, 43.9, 44.0, 44.0, 44.2, 44.9, 45.5, 46.4, 46.5, 47.5, 48.8, 49.2, 49.2, 50.1, 50.4, 50.4, 50.5, 50.9, 50.9, 52.1, 52.1, 52.1, 52.1, 52.1, 52.1, 53.4, 53.5, 53.8, 53.8, 54.7, 56.0, 56.0, 58.4, 58.5, 59.0, 59.0, 75.7 ] }, { "x": 4096.0, "output_throughput": { "median": 120.66, "min": 118.44, "max": 120.85, "n": 3 }, "median_ttft_ms": { "median": 589.45, "min": 587.65, "max": 591.46, "n": 3 }, "p99_ttft_ms": { "median": 3850.89, "min": 3091.9, "max": 4324.48, "n": 3 }, "median_tpot_ms": { "median": 53.76, "min": 47.83, "max": 54.23, "n": 3 }, "p99_tpot_ms": { "median": 104.06, "min": 96.84, "max": 108.21, "n": 3 }, "p99_itl_ms": { "median": 74.36, "min": 71.17, "max": 74.79, "n": 3 }, "median_itl_ms": { "median": 48.35, "min": 45.9, "max": 52.64, "n": 3 }, "goodput_rps": { "median": 1.62, "min": 1.45, "max": 1.68, "n": 3 }, "request_throughput": { "median": 1.89, "min": 1.85, "max": 1.89, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 85.0, "thermal_share": 0.022 }, "ttft_quantiles_ms": [ 65.0, 188.6, 247.7, 249.8, 251.5, 260.1, 263.0, 561.4, 563.0, 583.5, 584.9, 585.4, 585.7, 585.9, 586.2, 586.4, 586.7, 586.8, 587.0, 587.1, 587.4, 587.7, 588.0, 588.3, 588.7, 589.0, 589.5, 590.0, 590.2, 590.9, 591.5, 591.9, 592.8, 594.5, 595.3, 596.0, 596.3, 596.8, 597.4, 601.6, 602.8, 614.3, 656.3, 918.2, 1099.3, 2074.3, 2276.0, 2667.4, 3215.3, 3725.4, 4324.6 ], "tpot_quantiles_ms": [ 44.6, 44.6, 44.7, 44.8, 45.3, 46.4, 46.4, 46.5, 46.5, 47.2, 47.3, 47.3, 47.8, 47.8, 47.8, 47.9, 47.9, 48.4, 48.4, 48.6, 48.6, 48.7, 48.7, 48.7, 49.1, 49.1, 51.7, 53.2, 53.4, 53.4, 53.8, 53.8, 54.0, 54.4, 54.4, 55.6, 55.7, 56.0, 56.0, 56.2, 56.2, 58.8, 58.8, 59.7, 62.2, 62.4, 71.6, 78.3, 82.4, 100.4, 117.2 ] } ] }, "no-radix-cache": { "kv_pool_tokens": 238414, "points": [ { "x": 0.0, "output_throughput": { "median": 122.11, "min": 117.57, "max": 122.9, "n": 3 }, "median_ttft_ms": { "median": 564.63, "min": 504.58, "max": 585.69, "n": 3 }, "p99_ttft_ms": { "median": 762.84, "min": 760.84, "max": 1000.19, "n": 3 }, "median_tpot_ms": { "median": 55.28, "min": 54.73, "max": 59.67, "n": 3 }, "p99_tpot_ms": { "median": 76.62, "min": 73.97, "max": 78.42, "n": 3 }, "p99_itl_ms": { "median": 77.37, "min": 74.11, "max": 78.22, "n": 3 }, "median_itl_ms": { "median": 55.53, "min": 53.91, "max": 57.71, "n": 3 }, "goodput_rps": { "median": 1.84, "min": 1.79, "max": 1.92, "n": 3 }, "request_throughput": { "median": 1.91, "min": 1.84, "max": 1.92, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2496.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 207.0, 212.3, 214.3, 217.8, 264.0, 273.6, 287.4, 439.8, 452.0, 466.4, 493.8, 494.1, 499.1, 503.6, 503.8, 504.6, 525.8, 525.9, 544.0, 544.2, 562.0, 563.5, 564.5, 572.2, 573.3, 579.5, 580.5, 584.6, 585.2, 585.5, 586.1, 589.8, 590.9, 591.6, 592.1, 592.9, 720.6, 721.6, 723.6, 725.2, 725.7, 726.0, 728.2, 729.7, 730.6, 734.4, 735.6, 740.0, 759.4, 809.2, 1000.2 ], "tpot_quantiles_ms": [ 41.2, 44.8, 46.0, 46.0, 47.5, 47.5, 48.7, 48.7, 48.7, 48.7, 49.8, 49.8, 50.8, 51.3, 52.4, 52.5, 53.9, 53.9, 54.3, 54.3, 54.7, 54.7, 55.3, 55.3, 55.3, 55.3, 56.7, 58.3, 58.3, 59.2, 59.3, 59.7, 59.8, 62.4, 62.7, 63.3, 63.6, 64.2, 64.2, 67.3, 67.5, 67.6, 67.9, 68.7, 69.6, 69.7, 72.4, 73.9, 74.8, 76.6, 79.3 ] }, { "x": 256.0, "output_throughput": { "median": 122.58, "min": 122.18, "max": 130.86, "n": 3 }, "median_ttft_ms": { "median": 867.9, "min": 865.87, "max": 887.47, "n": 3 }, "p99_ttft_ms": { "median": 1016.19, "min": 900.2, "max": 1066.79, "n": 3 }, "median_tpot_ms": { "median": 50.82, "min": 49.1, "max": 50.99, "n": 3 }, "p99_tpot_ms": { "median": 70.74, "min": 67.6, "max": 78.73, "n": 3 }, "p99_itl_ms": { "median": 70.73, "min": 68.19, "max": 75.21, "n": 3 }, "median_itl_ms": { "median": 47.62, "min": 47.43, "max": 50.79, "n": 3 }, "goodput_rps": { "median": 1.88, "min": 1.73, "max": 1.92, "n": 3 }, "request_throughput": { "median": 1.92, "min": 1.91, "max": 2.04, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2496.0, "n": 3 }, "max_temp_c": 85.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 253.6, 269.6, 295.9, 304.7, 311.2, 313.3, 327.6, 586.0, 854.9, 856.5, 857.6, 858.6, 859.2, 860.1, 860.3, 861.4, 865.1, 865.6, 866.0, 866.7, 867.1, 867.2, 867.4, 867.7, 867.9, 868.2, 869.3, 869.7, 870.3, 874.1, 875.6, 880.0, 881.6, 882.1, 882.5, 882.7, 883.9, 884.6, 885.9, 886.7, 887.4, 887.8, 892.4, 893.5, 893.8, 894.5, 898.2, 898.4, 1015.8, 1066.7, 1066.8 ], "tpot_quantiles_ms": [ 40.0, 40.0, 40.3, 40.3, 41.1, 41.1, 41.2, 41.2, 41.3, 41.4, 42.1, 42.4, 42.4, 42.5, 42.5, 46.9, 46.9, 48.9, 48.9, 49.1, 49.1, 49.9, 50.0, 50.2, 50.2, 50.5, 50.8, 51.0, 51.0, 51.2, 51.8, 51.8, 52.9, 52.9, 54.9, 57.8, 58.8, 58.8, 60.2, 61.1, 61.1, 61.2, 61.2, 63.8, 63.9, 66.0, 66.0, 69.7, 71.3, 71.3, 79.8 ] }, { "x": 1024.0, "output_throughput": { "median": 102.13, "min": 99.76, "max": 102.21, "n": 3 }, "median_ttft_ms": { "median": 1675.28, "min": 1670.94, "max": 1689.6, "n": 3 }, "p99_ttft_ms": { "median": 1949.74, "min": 1940.19, "max": 2301.27, "n": 3 }, "median_tpot_ms": { "median": 52.78, "min": 49.54, "max": 54.89, "n": 3 }, "p99_tpot_ms": { "median": 81.1, "min": 76.92, "max": 82.24, "n": 3 }, "p99_itl_ms": { "median": 281.32, "min": 271.26, "max": 293.35, "n": 3 }, "median_itl_ms": { "median": 49.57, "min": 45.23, "max": 50.4, "n": 3 }, "goodput_rps": { "median": 0.22, "min": 0.19, "max": 0.25, "n": 3 }, "request_throughput": { "median": 1.6, "min": 1.56, "max": 1.6, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2408.0, "max": 2470.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.078 }, "ttft_quantiles_ms": [ 360.3, 374.0, 404.1, 407.9, 415.4, 417.5, 432.0, 750.5, 1662.1, 1662.4, 1663.0, 1664.0, 1665.3, 1667.0, 1667.6, 1667.8, 1668.0, 1668.4, 1670.6, 1672.2, 1673.0, 1673.6, 1674.9, 1675.2, 1676.2, 1679.3, 1679.8, 1681.3, 1682.6, 1685.4, 1686.0, 1687.3, 1688.5, 1689.5, 1690.8, 1692.6, 1693.4, 1694.7, 1882.1, 1889.0, 1893.6, 1898.9, 1905.2, 1907.7, 1908.6, 1921.6, 1959.7, 2022.9, 2023.1, 2215.5, 2447.0 ], "tpot_quantiles_ms": [ 42.4, 42.9, 42.9, 44.2, 45.9, 45.9, 46.3, 46.3, 46.7, 47.3, 47.3, 47.7, 47.7, 47.7, 49.4, 49.4, 49.5, 49.7, 49.7, 50.0, 51.8, 52.2, 52.2, 52.4, 52.8, 53.2, 53.5, 53.7, 53.7, 54.9, 55.1, 55.7, 55.9, 56.0, 56.8, 56.9, 57.0, 58.1, 58.5, 58.5, 62.4, 62.4, 66.0, 66.4, 66.9, 67.9, 69.8, 73.6, 76.3, 78.1, 86.1 ] }, { "x": 2048.0, "output_throughput": { "median": 79.07, "min": 77.83, "max": 79.98, "n": 3 }, "median_ttft_ms": { "median": 2577.17, "min": 2566.66, "max": 2578.04, "n": 3 }, "p99_ttft_ms": { "median": 3290.51, "min": 3285.73, "max": 3345.18, "n": 3 }, "median_tpot_ms": { "median": 65.37, "min": 62.1, "max": 65.5, "n": 3 }, "p99_tpot_ms": { "median": 100.75, "min": 97.98, "max": 101.82, "n": 3 }, "p99_itl_ms": { "median": 82.13, "min": 76.38, "max": 82.33, "n": 3 }, "median_itl_ms": { "median": 47.95, "min": 47.56, "max": 51.22, "n": 3 }, "goodput_rps": { "median": 0.14, "min": 0.13, "max": 0.16, "n": 3 }, "request_throughput": { "median": 1.24, "min": 1.22, "max": 1.25, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2405.0, "max": 2450.0, "n": 3 }, "max_temp_c": 88.0, "thermal_share": 0.175 }, "ttft_quantiles_ms": [ 521.5, 558.4, 559.6, 562.8, 570.2, 584.6, 596.9, 1846.5, 1855.5, 1857.3, 1858.1, 1859.6, 1866.6, 1870.8, 1876.2, 1877.6, 1880.3, 1883.9, 1885.0, 1886.1, 1890.5, 1892.4, 1901.0, 1911.1, 1912.7, 1925.1, 3231.6, 3241.3, 3241.8, 3242.1, 3242.5, 3246.0, 3246.5, 3247.8, 3253.3, 3258.9, 3261.3, 3262.5, 3268.5, 3272.6, 3273.8, 3277.2, 3279.6, 3282.0, 3284.4, 3284.5, 3285.7, 3290.7, 3292.3, 3342.2, 3345.2 ], "tpot_quantiles_ms": [ 45.5, 45.5, 45.6, 45.9, 46.5, 47.0, 47.2, 47.2, 47.5, 47.6, 47.8, 47.9, 48.2, 48.2, 49.8, 51.6, 53.5, 54.1, 54.4, 54.9, 55.4, 55.6, 56.6, 56.6, 62.3, 63.7, 67.6, 67.9, 68.4, 69.0, 69.2, 69.5, 69.7, 70.1, 70.1, 70.2, 72.0, 75.5, 76.4, 76.5, 76.9, 78.2, 78.6, 84.5, 88.4, 88.9, 90.0, 91.3, 96.0, 98.4, 105.8 ] }, { "x": 4096.0, "output_throughput": { "median": 49.73, "min": 48.83, "max": 49.84, "n": 3 }, "median_ttft_ms": { "median": 5208.24, "min": 4620.56, "max": 5224.42, "n": 3 }, "p99_ttft_ms": { "median": 6880.54, "min": 6878.23, "max": 7025.3, "n": 3 }, "median_tpot_ms": { "median": 83.16, "min": 81.98, "max": 93.85, "n": 3 }, "p99_tpot_ms": { "median": 152.42, "min": 151.36, "max": 160.26, "n": 3 }, "p99_itl_ms": { "median": 1380.6, "min": 1343.89, "max": 1382.7, "n": 3 }, "median_itl_ms": { "median": 55.13, "min": 54.94, "max": 60.39, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.78, "min": 0.76, "max": 0.78, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2375.5, "min": 2372.0, "max": 2379.0, "n": 3 }, "max_temp_c": 89.0, "thermal_share": 0.37 }, "ttft_quantiles_ms": [ 877.3, 922.7, 923.4, 929.8, 937.3, 949.5, 960.0, 2290.5, 2314.3, 2349.4, 2378.7, 2416.9, 2518.1, 3749.5, 3758.0, 3765.1, 3807.8, 3836.9, 3846.0, 3885.0, 3894.9, 3929.9, 3958.2, 3972.0, 4341.2, 5217.5, 5224.3, 5231.3, 5289.7, 5308.7, 5350.7, 5373.5, 5432.2, 5448.4, 5485.2, 5490.4, 5563.6, 6430.7, 6469.3, 6511.1, 6534.2, 6606.1, 6624.9, 6690.7, 6734.7, 6739.1, 6793.3, 6849.7, 6850.3, 6880.5, 7025.3 ], "tpot_quantiles_ms": [ 52.7, 53.4, 53.7, 53.9, 55.0, 55.3, 55.6, 56.0, 56.5, 60.9, 61.8, 62.9, 65.4, 68.7, 73.3, 74.2, 74.7, 75.9, 76.0, 76.7, 77.1, 77.9, 81.6, 82.0, 83.8, 88.0, 95.3, 97.4, 98.0, 98.6, 100.5, 100.9, 101.3, 102.4, 104.7, 105.2, 106.7, 110.8, 115.0, 121.3, 124.3, 124.9, 128.0, 130.6, 141.7, 145.1, 147.2, 149.0, 149.9, 152.0, 162.1 ] } ] } }, "rate": { "default": { "kv_pool_tokens": 238143, "points": [ { "x": 0.5, "output_throughput": { "median": 66.93, "min": 66.03, "max": 73.25, "n": 3 }, "median_ttft_ms": { "median": 476.31, "min": 476.01, "max": 476.65, "n": 3 }, "p99_ttft_ms": { "median": 658.8, "min": 652.09, "max": 963.89, "n": 3 }, "median_tpot_ms": { "median": 177.53, "min": 147.35, "max": 187.71, "n": 3 }, "p99_tpot_ms": { "median": 193.31, "min": 179.61, "max": 206.33, "n": 3 }, "goodput_rps": { "median": 0.07, "min": 0.0, "max": 0.1, "n": 3 }, "request_throughput": { "median": 0.52, "min": 0.52, "max": 0.57, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 43.6, 43.8, 275.6, 325.8, 353.0, 356.9, 367.0, 383.6, 387.9, 388.4, 390.2, 392.4, 400.7, 409.7, 413.3, 421.9, 435.1, 437.7, 442.2, 445.4, 450.6, 453.0, 461.2, 462.6, 470.6, 476.1, 478.6, 482.1, 485.5, 486.1, 505.7, 513.9, 520.3, 522.8, 533.7, 541.9, 554.1, 562.0, 567.6, 574.6, 579.4, 586.4, 593.9, 613.2, 627.1, 630.8, 633.7, 654.7, 672.1, 674.2, 1003.5 ], "tpot_quantiles_ms": [ 58.5, 67.8, 77.5, 84.4, 87.2, 99.8, 103.2, 109.5, 112.1, 114.1, 115.5, 127.4, 139.6, 144.9, 149.8, 150.5, 153.5, 154.8, 161.0, 161.8, 168.1, 168.1, 170.0, 170.5, 171.5, 173.6, 174.1, 174.9, 175.6, 176.4, 177.0, 177.8, 178.6, 178.7, 181.0, 183.3, 186.1, 187.6, 189.3, 190.4, 191.4, 192.0, 193.2, 194.0, 197.1, 199.4, 201.2, 202.2, 203.6, 204.7, 206.8 ] }, { "x": 1.0, "output_throughput": { "median": 99.22, "min": 92.62, "max": 99.25, "n": 3 }, "median_ttft_ms": { "median": 249.55, "min": 230.64, "max": 253.71, "n": 3 }, "p99_ttft_ms": { "median": 385.97, "min": 380.35, "max": 523.77, "n": 3 }, "median_tpot_ms": { "median": 192.29, "min": 167.81, "max": 197.63, "n": 3 }, "p99_tpot_ms": { "median": 200.9, "min": 178.98, "max": 208.85, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.78, "min": 0.72, "max": 0.78, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 78.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.3, 42.8, 62.5, 67.0, 104.6, 105.9, 111.5, 117.0, 124.0, 132.6, 143.6, 144.4, 158.0, 162.2, 179.8, 185.2, 187.6, 193.0, 196.9, 205.7, 210.9, 216.1, 229.1, 233.7, 240.7, 245.6, 250.7, 254.7, 256.7, 259.3, 269.0, 277.1, 282.6, 284.9, 290.5, 299.4, 301.8, 305.8, 307.8, 315.7, 326.7, 329.9, 339.0, 346.2, 350.2, 355.0, 363.1, 387.1, 396.3, 458.0, 549.1 ], "tpot_quantiles_ms": [ 84.2, 110.4, 127.9, 130.8, 139.4, 142.3, 142.5, 144.3, 152.7, 161.2, 162.2, 164.1, 166.6, 167.9, 168.8, 170.1, 171.0, 173.7, 174.8, 175.8, 177.2, 177.8, 178.9, 179.0, 181.7, 182.2, 187.3, 189.1, 191.3, 191.7, 191.9, 192.2, 194.1, 195.3, 196.7, 196.9, 196.9, 197.9, 198.9, 199.3, 199.3, 199.9, 200.7, 200.8, 202.9, 203.6, 205.0, 207.3, 208.4, 208.8, 208.9 ] }, { "x": 2.0, "output_throughput": { "median": 140.38, "min": 139.89, "max": 148.76, "n": 3 }, "median_ttft_ms": { "median": 427.19, "min": 420.72, "max": 731.61, "n": 3 }, "p99_ttft_ms": { "median": 916.83, "min": 867.0, "max": 1220.7, "n": 3 }, "median_tpot_ms": { "median": 282.33, "min": 281.97, "max": 287.17, "n": 3 }, "p99_tpot_ms": { "median": 298.56, "min": 296.53, "max": 326.71, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.1, "min": 1.09, "max": 1.16, "n": 3 }, "completed": [ 60, 60, 60 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.6, 65.2, 104.0, 118.9, 126.6, 148.8, 168.7, 184.4, 210.3, 238.1, 266.9, 271.0, 273.3, 287.2, 324.4, 340.6, 366.9, 420.6, 443.3, 484.0, 512.2, 531.5, 559.8, 581.3, 587.2, 595.0, 608.8, 611.3, 619.8, 626.6, 639.7, 652.0, 664.5, 676.6, 687.5, 697.8, 707.9, 719.3, 729.8, 749.7, 769.3, 777.0, 796.7, 836.8, 851.0, 863.6, 868.8, 925.9, 957.1, 1059.8, 1221.9 ], "tpot_quantiles_ms": [ 198.2, 225.5, 229.9, 231.3, 236.0, 237.9, 239.7, 243.0, 247.0, 250.4, 250.9, 254.3, 256.8, 259.1, 262.3, 263.5, 266.5, 267.1, 270.9, 274.0, 275.8, 277.8, 278.7, 280.0, 281.1, 283.4, 284.5, 286.0, 287.5, 287.8, 288.4, 289.5, 290.0, 291.5, 292.2, 293.8, 294.5, 295.1, 295.6, 296.1, 296.4, 296.5, 297.0, 297.6, 298.0, 298.6, 303.1, 310.2, 319.2, 325.3, 328.8 ] }, { "x": 4.0, "output_throughput": { "median": 240.31, "min": 239.51, "max": 246.34, "n": 3 }, "median_ttft_ms": { "median": 495.26, "min": 467.6, "max": 519.08, "n": 3 }, "p99_ttft_ms": { "median": 1516.17, "min": 998.46, "max": 1682.14, "n": 3 }, "median_tpot_ms": { "median": 355.58, "min": 347.87, "max": 362.97, "n": 3 }, "p99_tpot_ms": { "median": 374.45, "min": 368.29, "max": 377.3, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.88, "min": 1.87, "max": 1.92, "n": 3 }, "completed": [ 120, 120, 120 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 83.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.5, 68.2, 92.5, 111.2, 121.8, 138.6, 154.5, 172.3, 187.1, 196.1, 216.2, 228.4, 248.4, 268.4, 286.1, 292.3, 305.7, 314.0, 344.2, 358.9, 375.4, 400.3, 425.8, 439.9, 465.0, 492.0, 535.1, 597.3, 622.7, 665.1, 678.5, 692.5, 713.9, 739.0, 765.3, 779.3, 799.3, 809.8, 820.0, 831.1, 841.6, 868.0, 897.4, 923.5, 967.9, 1011.5, 1091.8, 1130.5, 1250.7, 1506.2, 1723.8 ], "tpot_quantiles_ms": [ 255.3, 273.7, 281.8, 287.0, 290.3, 293.3, 299.5, 304.1, 308.1, 312.6, 318.1, 321.0, 323.0, 328.0, 329.7, 333.7, 336.7, 340.1, 342.1, 344.5, 345.7, 347.6, 349.8, 351.8, 352.7, 354.3, 356.3, 358.5, 359.8, 362.6, 362.9, 363.9, 365.5, 365.8, 366.5, 366.8, 367.4, 367.9, 368.2, 368.8, 372.0, 372.6, 373.4, 373.9, 374.3, 375.5, 376.3, 376.5, 376.7, 376.8, 377.3 ] }, { "x": 8.0, "output_throughput": { "median": 407.28, "min": 401.74, "max": 407.65, "n": 3 }, "median_ttft_ms": { "median": 853.04, "min": 583.68, "max": 874.42, "n": 3 }, "p99_ttft_ms": { "median": 2946.56, "min": 2151.42, "max": 3396.28, "n": 3 }, "median_tpot_ms": { "median": 468.92, "min": 457.93, "max": 473.75, "n": 3 }, "p99_tpot_ms": { "median": 486.08, "min": 478.12, "max": 499.51, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 3.18, "min": 3.14, "max": 3.18, "n": 3 }, "completed": [ 240, 240, 240 ], "gpu": { "median_sm_mhz": { "median": 2470.0, "min": 2470.0, "max": 2470.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.0, 87.6, 109.3, 128.9, 142.7, 154.3, 174.9, 191.4, 221.6, 242.5, 257.4, 270.2, 296.3, 312.2, 333.3, 358.6, 369.7, 383.5, 393.8, 414.3, 438.6, 465.1, 515.6, 548.2, 625.6, 846.7, 969.7, 1023.6, 1052.1, 1106.9, 1149.2, 1184.1, 1209.3, 1241.2, 1274.2, 1330.4, 1374.0, 1445.4, 1526.7, 1583.6, 1658.0, 1740.8, 1821.9, 2001.2, 2140.3, 2356.3, 2578.1, 2715.5, 2856.4, 3037.5, 3709.9 ], "tpot_quantiles_ms": [ 353.3, 364.9, 365.7, 371.0, 374.6, 377.3, 381.8, 389.8, 395.7, 403.4, 403.6, 410.5, 415.6, 421.2, 423.9, 432.7, 437.7, 440.9, 445.1, 448.8, 452.3, 455.0, 458.4, 459.6, 462.8, 466.5, 469.0, 470.8, 473.3, 475.5, 476.7, 477.1, 477.2, 477.4, 477.5, 477.9, 478.1, 480.4, 481.3, 482.8, 484.0, 484.8, 485.5, 486.1, 487.5, 490.5, 495.4, 497.7, 498.5, 498.8, 499.5 ] }, { "x": 16.0, "output_throughput": { "median": 497.22, "min": 488.5, "max": 500.26, "n": 3 }, "median_ttft_ms": { "median": 790.85, "min": 784.82, "max": 1012.68, "n": 3 }, "p99_ttft_ms": { "median": 55640.39, "min": 54041.76, "max": 57290.85, "n": 3 }, "median_tpot_ms": { "median": 571.79, "min": 568.63, "max": 577.1, "n": 3 }, "p99_tpot_ms": { "median": 620.14, "min": 614.18, "max": 630.29, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 3.88, "min": 3.82, "max": 3.91, "n": 3 }, "completed": [ 480, 480, 480 ], "gpu": { "median_sm_mhz": { "median": 2444.0, "min": 2444.0, "max": 2450.0, "n": 3 }, "max_temp_c": 89.0, "thermal_share": 0.073 }, "ttft_quantiles_ms": [ 42.0, 105.6, 127.8, 143.0, 160.1, 177.6, 196.1, 207.3, 222.2, 245.7, 272.8, 292.8, 313.0, 337.0, 366.6, 389.4, 410.7, 434.5, 456.3, 485.4, 518.8, 557.7, 613.5, 686.3, 763.0, 935.1, 1202.3, 1526.0, 1890.8, 2089.4, 2403.7, 2801.7, 3015.3, 3158.4, 3397.7, 3658.4, 4104.4, 4930.1, 5737.1, 6601.1, 46394.2, 48569.6, 49398.1, 50759.3, 51279.1, 52519.5, 53538.3, 54183.5, 54906.9, 55820.3, 57590.7 ], "tpot_quantiles_ms": [ 296.2, 309.2, 319.6, 329.2, 334.5, 346.8, 350.0, 367.0, 375.6, 378.6, 394.1, 508.6, 518.4, 520.0, 522.1, 530.0, 538.4, 540.1, 542.5, 548.4, 549.7, 552.1, 559.6, 560.8, 566.6, 568.6, 575.7, 580.5, 581.7, 587.7, 590.5, 592.4, 597.9, 599.1, 602.0, 603.0, 607.1, 608.5, 609.4, 611.4, 612.6, 613.1, 615.0, 617.0, 618.2, 619.3, 621.3, 622.4, 624.0, 627.9, 630.3 ] } ] }, "fp8": { "kv_pool_tokens": 1082427, "points": [ { "x": 0.5, "output_throughput": { "median": 76.41, "min": 73.91, "max": 94.19, "n": 3 }, "median_ttft_ms": { "median": 208.44, "min": 206.02, "max": 216.72, "n": 3 }, "p99_ttft_ms": { "median": 284.93, "min": 253.26, "max": 321.71, "n": 3 }, "median_tpot_ms": { "median": 59.78, "min": 50.57, "max": 64.56, "n": 3 }, "p99_tpot_ms": { "median": 71.76, "min": 66.31, "max": 78.52, "n": 3 }, "goodput_rps": { "median": 0.6, "min": 0.58, "max": 0.74, "n": 3 }, "request_throughput": { "median": 0.6, "min": 0.58, "max": 0.74, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 43.1, 44.7, 92.1, 140.0, 141.4, 142.7, 156.1, 156.9, 160.1, 162.2, 167.0, 168.8, 173.5, 176.9, 179.3, 183.5, 190.2, 192.9, 194.3, 198.1, 199.0, 199.5, 202.4, 204.2, 206.8, 209.5, 210.8, 212.6, 213.9, 216.3, 217.1, 219.6, 223.0, 224.6, 227.2, 228.7, 230.7, 233.3, 235.2, 238.2, 241.7, 243.4, 243.9, 246.1, 248.0, 250.3, 253.1, 254.9, 272.6, 290.5, 325.0 ], "tpot_quantiles_ms": [ 18.0, 18.4, 18.6, 21.6, 22.7, 23.6, 25.4, 29.1, 30.4, 32.6, 33.2, 37.4, 44.9, 46.5, 46.8, 49.0, 50.5, 51.5, 51.6, 51.8, 51.9, 52.4, 53.4, 57.6, 58.1, 58.8, 59.7, 60.0, 60.1, 61.1, 61.6, 61.7, 62.5, 62.9, 64.0, 64.6, 65.4, 66.1, 66.2, 67.1, 68.2, 68.2, 68.4, 69.4, 70.1, 71.5, 72.0, 72.7, 75.8, 76.8, 78.8 ] }, { "x": 1.0, "output_throughput": { "median": 132.45, "min": 131.88, "max": 147.83, "n": 3 }, "median_ttft_ms": { "median": 137.75, "min": 128.79, "max": 156.0, "n": 3 }, "p99_ttft_ms": { "median": 217.66, "min": 215.93, "max": 238.04, "n": 3 }, "median_tpot_ms": { "median": 77.97, "min": 70.74, "max": 86.4, "n": 3 }, "p99_tpot_ms": { "median": 85.96, "min": 81.56, "max": 90.1, "n": 3 }, "goodput_rps": { "median": 1.03, "min": 1.03, "max": 1.15, "n": 3 }, "request_throughput": { "median": 1.03, "min": 1.03, "max": 1.15, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 43.2, 43.8, 47.3, 72.4, 76.4, 84.0, 87.0, 91.9, 93.9, 96.5, 99.6, 102.4, 105.2, 107.6, 108.8, 113.1, 114.3, 116.8, 117.3, 119.4, 121.2, 122.0, 127.2, 129.4, 136.1, 139.2, 141.1, 142.9, 145.2, 146.6, 149.8, 149.9, 150.9, 151.2, 154.6, 158.2, 159.0, 162.7, 168.1, 170.0, 170.5, 171.0, 175.7, 185.0, 186.3, 189.4, 191.3, 198.3, 210.9, 219.1, 255.9 ], "tpot_quantiles_ms": [ 30.3, 33.7, 38.7, 41.5, 48.9, 50.3, 51.0, 51.7, 54.8, 58.6, 59.3, 59.4, 66.6, 68.2, 69.2, 72.3, 73.5, 74.5, 75.3, 75.7, 76.0, 76.9, 77.5, 78.4, 78.6, 79.2, 79.5, 79.8, 79.9, 80.5, 81.1, 81.3, 81.6, 81.6, 82.5, 84.2, 84.5, 84.6, 84.8, 85.2, 85.5, 85.5, 86.3, 87.4, 87.9, 89.0, 89.8, 89.9, 90.0, 90.0, 90.1 ] }, { "x": 2.0, "output_throughput": { "median": 210.14, "min": 202.4, "max": 222.18, "n": 3 }, "median_ttft_ms": { "median": 215.3, "min": 212.09, "max": 218.37, "n": 3 }, "p99_ttft_ms": { "median": 399.71, "min": 386.15, "max": 413.21, "n": 3 }, "median_tpot_ms": { "median": 127.43, "min": 107.47, "max": 128.5, "n": 3 }, "p99_tpot_ms": { "median": 140.01, "min": 127.39, "max": 143.66, "n": 3 }, "goodput_rps": { "median": 0.17, "min": 0.16, "max": 0.4, "n": 3 }, "request_throughput": { "median": 1.64, "min": 1.58, "max": 1.74, "n": 3 }, "completed": [ 60, 60, 60 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 43.0, 45.7, 56.4, 69.3, 75.5, 89.5, 98.0, 108.9, 111.8, 118.8, 125.9, 134.4, 138.2, 142.3, 148.4, 153.6, 159.3, 165.7, 167.2, 179.5, 181.8, 189.6, 191.1, 205.9, 210.2, 212.6, 226.4, 240.8, 256.2, 260.5, 264.9, 272.2, 282.7, 289.8, 293.0, 304.7, 311.9, 315.8, 323.2, 326.7, 337.4, 345.3, 350.8, 351.2, 356.0, 360.5, 364.5, 372.4, 376.1, 393.3, 418.8 ], "tpot_quantiles_ms": [ 55.3, 75.5, 87.8, 90.8, 93.8, 95.6, 97.3, 98.6, 100.5, 102.0, 102.9, 104.8, 105.7, 106.9, 107.5, 107.9, 110.8, 112.6, 113.1, 115.0, 115.8, 117.1, 117.9, 118.7, 119.1, 120.3, 121.9, 122.0, 124.2, 125.2, 126.1, 126.9, 127.1, 127.4, 127.8, 129.6, 131.5, 132.7, 133.9, 135.1, 136.6, 137.9, 138.1, 138.9, 139.5, 139.8, 140.6, 142.0, 142.9, 143.4, 143.7 ] }, { "x": 4.0, "output_throughput": { "median": 361.67, "min": 342.05, "max": 364.83, "n": 3 }, "median_ttft_ms": { "median": 271.25, "min": 270.71, "max": 276.52, "n": 3 }, "p99_ttft_ms": { "median": 513.51, "min": 447.34, "max": 545.18, "n": 3 }, "median_tpot_ms": { "median": 173.98, "min": 168.55, "max": 178.41, "n": 3 }, "p99_tpot_ms": { "median": 192.73, "min": 190.55, "max": 196.12, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 2.83, "min": 2.67, "max": 2.85, "n": 3 }, "completed": [ 120, 120, 120 ], "gpu": { "median_sm_mhz": { "median": 2496.0, "min": 2496.0, "max": 2496.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.0, 54.4, 61.0, 68.6, 77.9, 81.4, 102.0, 111.1, 121.8, 128.1, 142.4, 158.1, 174.5, 181.0, 184.1, 189.5, 192.8, 198.0, 204.1, 217.4, 220.4, 228.1, 237.7, 249.4, 267.8, 272.7, 277.8, 287.6, 295.0, 306.1, 313.0, 328.8, 333.9, 337.9, 346.5, 351.8, 357.2, 362.6, 376.4, 380.9, 389.2, 394.5, 397.9, 404.5, 412.3, 420.9, 428.7, 448.2, 471.4, 512.7, 582.6 ], "tpot_quantiles_ms": [ 97.7, 117.5, 122.2, 125.4, 129.1, 131.9, 134.1, 138.0, 139.3, 141.5, 144.9, 147.4, 150.4, 151.1, 154.1, 156.6, 158.6, 161.3, 163.3, 165.2, 166.6, 168.5, 169.8, 171.8, 173.8, 175.0, 176.1, 177.5, 179.5, 180.6, 182.8, 183.9, 185.5, 186.7, 188.3, 189.0, 189.5, 189.9, 190.0, 190.1, 190.2, 190.4, 190.5, 191.2, 191.9, 192.6, 192.9, 195.0, 195.6, 195.9, 196.2 ] }, { "x": 8.0, "output_throughput": { "median": 635.69, "min": 610.15, "max": 639.82, "n": 3 }, "median_ttft_ms": { "median": 303.85, "min": 297.02, "max": 314.2, "n": 3 }, "p99_ttft_ms": { "median": 740.29, "min": 720.04, "max": 892.46, "n": 3 }, "median_tpot_ms": { "median": 246.76, "min": 238.23, "max": 247.66, "n": 3 }, "p99_tpot_ms": { "median": 257.07, "min": 253.63, "max": 258.96, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 4.97, "min": 4.77, "max": 5.0, "n": 3 }, "completed": [ 240, 240, 240 ], "gpu": { "median_sm_mhz": { "median": 2476.0, "min": 2476.0, "max": 2476.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.7, 56.5, 63.5, 77.5, 89.2, 100.4, 110.4, 118.5, 132.3, 143.2, 154.8, 162.8, 174.3, 183.4, 190.4, 198.6, 212.5, 219.4, 230.0, 238.2, 246.5, 254.1, 265.6, 276.8, 288.7, 303.4, 324.4, 344.6, 364.0, 381.1, 393.9, 404.3, 412.0, 430.6, 441.8, 454.2, 466.6, 478.7, 493.0, 499.2, 508.4, 519.3, 536.0, 551.2, 564.1, 587.5, 619.4, 657.3, 688.6, 749.6, 923.3 ], "tpot_quantiles_ms": [ 154.0, 163.9, 167.8, 170.8, 173.8, 178.1, 181.8, 185.1, 190.3, 194.1, 197.0, 201.3, 204.3, 208.3, 212.7, 215.5, 219.9, 222.9, 225.4, 228.4, 229.7, 233.7, 237.2, 239.8, 241.1, 244.1, 245.3, 247.2, 248.4, 249.0, 250.0, 250.8, 251.2, 251.7, 252.4, 252.7, 253.0, 253.2, 253.3, 253.8, 254.7, 255.2, 255.8, 256.2, 256.4, 256.5, 256.7, 256.9, 257.7, 258.5, 259.2 ] }, { "x": 16.0, "output_throughput": { "median": 1057.72, "min": 1012.6, "max": 1060.24, "n": 3 }, "median_ttft_ms": { "median": 382.32, "min": 374.39, "max": 382.97, "n": 3 }, "p99_ttft_ms": { "median": 1984.04, "min": 1439.8, "max": 2314.69, "n": 3 }, "median_tpot_ms": { "median": 339.84, "min": 337.37, "max": 342.32, "n": 3 }, "p99_tpot_ms": { "median": 352.51, "min": 350.02, "max": 359.82, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 8.26, "min": 7.91, "max": 8.28, "n": 3 }, "completed": [ 480, 480, 480 ], "gpu": { "median_sm_mhz": { "median": 2470.0, "min": 2470.0, "max": 2470.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.041 }, "ttft_quantiles_ms": [ 42.2, 62.6, 78.4, 86.4, 97.8, 105.7, 115.4, 127.3, 135.9, 148.6, 160.5, 179.0, 187.2, 198.8, 210.7, 219.3, 233.6, 243.1, 251.2, 263.2, 282.3, 295.5, 312.5, 330.8, 352.1, 379.6, 487.6, 585.3, 650.8, 716.0, 774.3, 827.1, 869.0, 943.1, 987.3, 1031.2, 1068.0, 1109.5, 1156.1, 1211.4, 1257.8, 1302.5, 1350.4, 1402.7, 1465.6, 1545.0, 1639.2, 1748.8, 1850.0, 2001.2, 2548.6 ], "tpot_quantiles_ms": [ 219.5, 223.7, 229.4, 233.2, 238.6, 244.3, 248.2, 251.8, 256.7, 261.1, 267.1, 272.1, 276.5, 281.5, 287.1, 291.3, 298.0, 302.1, 307.1, 312.2, 318.7, 322.0, 328.6, 333.2, 336.6, 339.2, 342.0, 343.1, 344.3, 345.3, 346.5, 347.2, 348.1, 348.6, 348.8, 349.1, 349.4, 349.7, 350.0, 350.8, 351.6, 351.8, 352.2, 353.5, 355.3, 356.4, 357.4, 358.5, 359.1, 359.4, 360.0 ] } ] }, "default-kv-max": { "kv_pool_tokens": 387239, "points": [ { "x": 0.5, "output_throughput": { "median": 67.33, "min": 67.08, "max": 74.16, "n": 3 }, "median_ttft_ms": { "median": 477.7, "min": 463.89, "max": 491.66, "n": 3 }, "p99_ttft_ms": { "median": 670.64, "min": 663.97, "max": 850.03, "n": 3 }, "median_tpot_ms": { "median": 172.41, "min": 146.41, "max": 180.66, "n": 3 }, "p99_tpot_ms": { "median": 187.05, "min": 176.81, "max": 199.97, "n": 3 }, "p99_itl_ms": { "median": 440.95, "min": 432.18, "max": 695.54, "n": 3 }, "median_itl_ms": { "median": 141.6, "min": 123.81, "max": 159.32, "n": 3 }, "goodput_rps": { "median": 0.08, "min": 0.0, "max": 0.1, "n": 3 }, "request_throughput": { "median": 0.53, "min": 0.52, "max": 0.58, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.3, 42.9, 308.3, 333.1, 342.6, 351.1, 359.6, 370.1, 383.2, 385.9, 390.5, 391.2, 403.1, 406.1, 412.2, 414.5, 416.5, 419.5, 430.7, 432.7, 449.8, 453.0, 462.6, 465.2, 468.3, 480.7, 484.8, 489.4, 493.9, 502.1, 505.6, 508.0, 518.9, 522.1, 529.9, 535.9, 538.8, 544.8, 550.7, 558.8, 569.0, 572.5, 590.4, 597.2, 601.7, 616.0, 620.1, 664.8, 674.2, 728.1, 850.0 ], "tpot_quantiles_ms": [ 57.1, 60.3, 75.8, 80.3, 84.9, 97.0, 100.4, 104.4, 106.8, 109.3, 112.2, 115.0, 137.5, 141.7, 144.1, 145.8, 148.7, 149.7, 156.9, 159.4, 160.5, 163.4, 165.8, 167.3, 167.7, 168.2, 169.4, 170.4, 170.7, 171.6, 172.4, 173.3, 174.2, 175.1, 175.7, 176.9, 177.3, 180.8, 181.7, 182.5, 184.9, 184.9, 186.3, 187.2, 189.5, 192.7, 193.2, 195.5, 196.9, 198.4, 200.4 ] }, { "x": 1.0, "output_throughput": { "median": 99.55, "min": 95.37, "max": 101.34, "n": 3 }, "median_ttft_ms": { "median": 246.12, "min": 244.7, "max": 261.53, "n": 3 }, "p99_ttft_ms": { "median": 425.72, "min": 420.91, "max": 447.47, "n": 3 }, "median_tpot_ms": { "median": 185.05, "min": 161.95, "max": 188.58, "n": 3 }, "p99_tpot_ms": { "median": 190.79, "min": 174.81, "max": 196.75, "n": 3 }, "p99_itl_ms": { "median": 264.29, "min": 229.45, "max": 303.58, "n": 3 }, "median_itl_ms": { "median": 177.58, "min": 168.81, "max": 186.51, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.78, "min": 0.75, "max": 0.79, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.1, 41.4, 69.0, 81.7, 86.5, 100.6, 105.1, 114.5, 133.0, 135.1, 146.2, 147.1, 163.2, 174.6, 191.5, 198.2, 200.1, 210.6, 216.4, 218.9, 226.1, 235.1, 235.4, 238.0, 242.5, 252.0, 253.0, 254.4, 268.7, 277.8, 288.6, 290.8, 295.9, 299.0, 309.8, 316.2, 324.5, 328.1, 338.0, 342.3, 344.0, 344.9, 346.8, 350.1, 361.0, 375.1, 384.3, 395.0, 407.0, 427.2, 463.3 ], "tpot_quantiles_ms": [ 83.3, 105.1, 123.4, 126.5, 126.8, 141.0, 141.1, 144.0, 150.8, 151.6, 155.8, 156.7, 159.4, 161.6, 162.0, 164.3, 167.0, 167.8, 169.4, 171.8, 172.1, 172.5, 172.8, 173.2, 174.5, 174.5, 174.9, 178.6, 181.0, 181.8, 183.4, 183.4, 183.6, 186.5, 186.8, 187.2, 187.7, 188.3, 188.5, 188.9, 189.6, 189.8, 189.9, 190.5, 190.9, 191.3, 192.9, 194.9, 196.0, 196.5, 196.8 ] }, { "x": 2.0, "output_throughput": { "median": 145.77, "min": 144.88, "max": 150.5, "n": 3 }, "median_ttft_ms": { "median": 396.21, "min": 352.05, "max": 400.77, "n": 3 }, "p99_ttft_ms": { "median": 863.19, "min": 836.96, "max": 931.14, "n": 3 }, "median_tpot_ms": { "median": 272.33, "min": 250.28, "max": 274.79, "n": 3 }, "p99_tpot_ms": { "median": 281.84, "min": 268.6, "max": 291.64, "n": 3 }, "p99_itl_ms": { "median": 1103.07, "min": 999.72, "max": 1353.5, "n": 3 }, "median_itl_ms": { "median": 230.69, "min": 217.01, "max": 240.31, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.14, "min": 1.13, "max": 1.18, "n": 3 }, "completed": [ 60, 60, 60 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 40.3, 54.9, 81.3, 81.7, 106.1, 116.3, 130.3, 140.1, 151.7, 166.4, 176.8, 185.7, 192.6, 205.4, 216.2, 230.5, 245.6, 252.4, 263.1, 280.3, 283.6, 323.1, 341.2, 349.0, 355.7, 388.6, 410.1, 458.2, 505.4, 516.7, 540.0, 565.1, 580.4, 589.3, 596.6, 621.9, 623.8, 638.8, 667.2, 670.4, 693.3, 699.1, 704.3, 714.1, 724.3, 748.3, 760.9, 773.9, 790.1, 843.9, 952.0 ], "tpot_quantiles_ms": [ 183.7, 211.0, 216.7, 219.4, 223.3, 225.2, 226.8, 230.3, 234.3, 236.4, 237.9, 241.2, 241.7, 243.1, 244.9, 246.1, 247.8, 250.5, 251.7, 254.3, 255.3, 256.8, 258.9, 260.9, 261.0, 262.3, 263.8, 264.8, 266.0, 267.9, 268.3, 268.4, 271.0, 272.7, 273.5, 274.8, 275.9, 277.7, 278.9, 280.8, 281.1, 281.2, 281.4, 281.7, 282.4, 283.6, 287.6, 289.9, 290.6, 291.4, 291.6 ] }, { "x": 4.0, "output_throughput": { "median": 249.15, "min": 242.19, "max": 250.44, "n": 3 }, "median_ttft_ms": { "median": 505.8, "min": 497.19, "max": 532.03, "n": 3 }, "p99_ttft_ms": { "median": 1085.07, "min": 1023.71, "max": 1280.22, "n": 3 }, "median_tpot_ms": { "median": 349.68, "min": 343.75, "max": 351.54, "n": 3 }, "p99_tpot_ms": { "median": 363.92, "min": 363.32, "max": 365.1, "n": 3 }, "p99_itl_ms": { "median": 1820.47, "min": 1645.42, "max": 2020.35, "n": 3 }, "median_itl_ms": { "median": 278.26, "min": 270.92, "max": 278.89, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 1.95, "min": 1.89, "max": 1.96, "n": 3 }, "completed": [ 120, 120, 120 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 40.6, 75.0, 92.0, 105.1, 113.7, 124.4, 136.6, 148.3, 170.2, 193.1, 204.6, 218.9, 238.1, 251.6, 277.5, 290.2, 300.8, 319.8, 328.7, 337.9, 351.4, 367.6, 391.2, 431.6, 472.2, 498.0, 543.4, 590.7, 623.9, 641.7, 668.9, 687.9, 695.6, 706.4, 724.9, 739.3, 746.9, 762.8, 775.0, 797.3, 809.6, 822.1, 845.5, 864.8, 887.6, 900.5, 920.8, 944.4, 991.3, 1029.8, 1354.2 ], "tpot_quantiles_ms": [ 251.2, 268.6, 271.7, 276.7, 279.9, 284.5, 290.7, 294.9, 299.4, 304.4, 308.2, 310.7, 315.1, 318.5, 322.2, 325.5, 328.8, 333.4, 334.6, 337.1, 337.9, 340.0, 341.8, 343.8, 347.1, 348.3, 349.2, 351.1, 352.0, 353.1, 354.9, 355.8, 356.1, 357.1, 358.1, 359.4, 360.3, 361.1, 361.4, 362.4, 362.6, 362.7, 362.8, 362.9, 363.0, 363.3, 363.5, 363.8, 364.1, 364.7, 365.3 ] }, { "x": 8.0, "output_throughput": { "median": 408.26, "min": 406.87, "max": 410.44, "n": 3 }, "median_ttft_ms": { "median": 628.53, "min": 579.15, "max": 722.09, "n": 3 }, "p99_ttft_ms": { "median": 1846.75, "min": 1604.86, "max": 2202.15, "n": 3 }, "median_tpot_ms": { "median": 461.01, "min": 452.69, "max": 464.27, "n": 3 }, "p99_tpot_ms": { "median": 480.8, "min": 473.92, "max": 483.38, "n": 3 }, "p99_itl_ms": { "median": 4802.53, "min": 963.26, "max": 5573.39, "n": 3 }, "median_itl_ms": { "median": 360.82, "min": 351.56, "max": 362.02, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 3.19, "min": 3.18, "max": 3.21, "n": 3 }, "completed": [ 240, 240, 240 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 83.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 39.5, 86.6, 108.7, 121.2, 131.9, 145.8, 161.6, 171.3, 190.1, 201.7, 219.9, 245.1, 255.7, 274.4, 294.9, 314.4, 325.5, 337.1, 345.2, 362.6, 379.2, 403.3, 448.4, 479.7, 532.2, 643.3, 772.4, 857.7, 883.8, 930.4, 963.0, 992.8, 1021.1, 1040.6, 1068.2, 1088.4, 1109.9, 1162.1, 1181.5, 1212.0, 1247.0, 1284.5, 1313.1, 1364.4, 1396.9, 1448.9, 1508.5, 1586.3, 1681.9, 1884.0, 2299.8 ], "tpot_quantiles_ms": [ 347.3, 364.6, 369.7, 373.9, 378.2, 380.8, 388.0, 392.3, 397.7, 402.7, 405.4, 412.4, 416.8, 420.0, 424.9, 429.2, 432.6, 437.4, 440.6, 444.9, 447.1, 451.4, 453.5, 456.3, 457.9, 459.3, 461.4, 462.1, 464.2, 465.5, 467.2, 468.2, 468.5, 469.9, 471.6, 472.0, 473.0, 473.4, 475.5, 477.6, 478.4, 478.7, 479.2, 479.4, 479.9, 480.2, 480.5, 480.9, 482.3, 482.8, 483.7 ] }, { "x": 16.0, "output_throughput": { "median": 664.28, "min": 651.52, "max": 666.35, "n": 3 }, "median_ttft_ms": { "median": 876.78, "min": 805.2, "max": 885.98, "n": 3 }, "p99_ttft_ms": { "median": 5920.88, "min": 5612.04, "max": 6793.1, "n": 3 }, "median_tpot_ms": { "median": 609.48, "min": 604.38, "max": 610.71, "n": 3 }, "p99_tpot_ms": { "median": 627.73, "min": 621.51, "max": 632.28, "n": 3 }, "p99_itl_ms": { "median": 1216.04, "min": 846.42, "max": 1331.35, "n": 3 }, "median_itl_ms": { "median": 452.72, "min": 452.47, "max": 453.41, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 5.19, "min": 5.09, "max": 5.21, "n": 3 }, "completed": [ 480, 480, 480 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2405.0, "max": 2405.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.019 }, "ttft_quantiles_ms": [ 41.3, 105.0, 129.8, 142.0, 154.8, 166.1, 179.2, 193.9, 211.8, 235.0, 254.3, 271.8, 288.2, 318.2, 334.5, 352.0, 368.3, 391.2, 409.9, 426.0, 449.7, 469.8, 495.4, 528.5, 592.5, 792.4, 1268.9, 1665.0, 1871.0, 2090.6, 2348.7, 2484.1, 2667.3, 2883.6, 3116.2, 3299.2, 3480.3, 3605.0, 3710.1, 3936.0, 4117.5, 4342.3, 4600.1, 4853.5, 5023.7, 5221.3, 5377.4, 5670.9, 6013.4, 6425.8, 7080.7 ], "tpot_quantiles_ms": [ 452.9, 458.4, 465.2, 468.4, 475.2, 478.5, 485.3, 495.3, 504.1, 505.3, 514.2, 520.7, 525.1, 534.3, 535.2, 544.3, 550.9, 555.1, 564.3, 569.9, 573.5, 582.9, 589.3, 596.4, 604.4, 609.5, 612.6, 613.7, 615.1, 616.0, 617.1, 617.7, 618.4, 618.6, 619.3, 619.7, 620.1, 620.3, 620.6, 621.0, 621.7, 622.7, 624.1, 625.0, 626.0, 626.4, 627.0, 627.7, 628.8, 630.3, 632.6 ] } ] }, "fp8-kv-max": { "kv_pool_tokens": 1378624, "points": [ { "x": 0.5, "output_throughput": { "median": 76.62, "min": 73.92, "max": 95.33, "n": 3 }, "median_ttft_ms": { "median": 201.95, "min": 192.75, "max": 213.03, "n": 3 }, "p99_ttft_ms": { "median": 278.28, "min": 263.21, "max": 294.5, "n": 3 }, "median_tpot_ms": { "median": 56.73, "min": 47.23, "max": 62.7, "n": 3 }, "p99_tpot_ms": { "median": 69.41, "min": 65.03, "max": 74.8, "n": 3 }, "p99_itl_ms": { "median": 191.67, "min": 180.91, "max": 197.72, "n": 3 }, "median_itl_ms": { "median": 53.0, "min": 39.55, "max": 58.0, "n": 3 }, "goodput_rps": { "median": 0.6, "min": 0.58, "max": 0.74, "n": 3 }, "request_throughput": { "median": 0.6, "min": 0.58, "max": 0.74, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 40.9, 41.3, 88.8, 136.4, 138.6, 141.2, 150.5, 154.5, 156.9, 160.2, 163.5, 164.7, 173.2, 174.1, 177.1, 178.6, 181.2, 182.9, 187.4, 191.0, 191.7, 192.1, 194.1, 196.1, 198.7, 202.2, 204.8, 207.0, 211.3, 212.3, 214.1, 215.8, 221.7, 224.7, 225.6, 226.7, 228.3, 228.9, 232.1, 234.0, 236.4, 236.8, 237.5, 238.7, 240.5, 241.6, 245.3, 249.2, 261.3, 268.0, 306.4 ], "tpot_quantiles_ms": [ 18.0, 18.0, 18.3, 21.5, 22.3, 23.1, 24.8, 27.4, 29.2, 31.0, 31.9, 34.9, 43.5, 44.6, 45.4, 46.0, 47.6, 48.0, 49.4, 49.6, 50.2, 51.3, 52.0, 52.9, 55.2, 55.6, 56.2, 58.0, 58.3, 59.4, 59.6, 59.9, 60.4, 61.4, 62.1, 62.7, 63.3, 64.1, 64.7, 64.9, 65.1, 65.4, 66.4, 67.4, 67.8, 68.7, 69.5, 71.3, 71.8, 73.1, 75.1 ] }, { "x": 1.0, "output_throughput": { "median": 133.53, "min": 132.4, "max": 149.98, "n": 3 }, "median_ttft_ms": { "median": 125.95, "min": 125.57, "max": 143.14, "n": 3 }, "p99_ttft_ms": { "median": 200.54, "min": 191.53, "max": 209.85, "n": 3 }, "median_tpot_ms": { "median": 74.8, "min": 68.07, "max": 83.56, "n": 3 }, "p99_tpot_ms": { "median": 82.13, "min": 78.55, "max": 86.91, "n": 3 }, "p99_itl_ms": { "median": 120.7, "min": 110.97, "max": 125.4, "n": 3 }, "median_itl_ms": { "median": 71.81, "min": 64.05, "max": 81.59, "n": 3 }, "goodput_rps": { "median": 1.04, "min": 1.03, "max": 1.17, "n": 3 }, "request_throughput": { "median": 1.04, "min": 1.03, "max": 1.17, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 39.7, 40.9, 41.5, 45.7, 62.6, 73.2, 75.2, 76.6, 77.0, 81.4, 87.4, 91.7, 97.2, 98.7, 102.4, 102.7, 104.4, 109.1, 110.2, 115.0, 118.6, 119.1, 122.5, 125.5, 126.5, 126.7, 130.3, 131.5, 134.3, 135.5, 136.2, 139.2, 146.0, 147.0, 147.1, 148.4, 149.2, 149.6, 157.6, 158.3, 161.7, 162.8, 170.8, 174.3, 178.8, 182.1, 187.5, 189.4, 192.5, 205.5, 210.0 ], "tpot_quantiles_ms": [ 29.6, 31.8, 37.0, 39.9, 44.3, 45.4, 47.6, 48.4, 51.3, 56.5, 57.3, 57.5, 64.5, 65.1, 66.8, 69.3, 70.6, 71.5, 71.9, 72.5, 73.0, 74.0, 74.3, 74.8, 75.4, 76.1, 76.5, 76.8, 77.0, 77.2, 78.0, 78.2, 78.6, 79.4, 79.5, 80.3, 80.8, 81.2, 81.4, 81.5, 82.3, 82.4, 83.5, 84.1, 84.8, 85.8, 86.3, 86.5, 86.8, 86.9, 86.9 ] }, { "x": 2.0, "output_throughput": { "median": 211.79, "min": 204.66, "max": 223.91, "n": 3 }, "median_ttft_ms": { "median": 222.12, "min": 210.59, "max": 223.19, "n": 3 }, "p99_ttft_ms": { "median": 394.11, "min": 380.94, "max": 394.86, "n": 3 }, "median_tpot_ms": { "median": 124.85, "min": 104.34, "max": 124.96, "n": 3 }, "p99_tpot_ms": { "median": 137.65, "min": 123.52, "max": 139.58, "n": 3 }, "p99_itl_ms": { "median": 365.85, "min": 357.32, "max": 376.84, "n": 3 }, "median_itl_ms": { "median": 105.89, "min": 94.05, "max": 111.92, "n": 3 }, "goodput_rps": { "median": 0.26, "min": 0.22, "max": 0.53, "n": 3 }, "request_throughput": { "median": 1.65, "min": 1.6, "max": 1.75, "n": 3 }, "completed": [ 60, 60, 60 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 41.2, 46.7, 61.4, 74.2, 80.8, 85.4, 92.9, 110.7, 120.5, 130.3, 132.7, 136.4, 142.0, 145.8, 149.7, 151.6, 164.8, 165.9, 175.5, 179.6, 181.0, 191.0, 195.6, 203.7, 209.2, 217.8, 230.6, 240.2, 244.5, 252.1, 258.6, 261.7, 267.0, 270.1, 271.2, 280.1, 284.4, 289.3, 294.2, 295.5, 301.2, 305.1, 320.1, 324.6, 335.8, 339.3, 341.9, 354.6, 357.4, 375.9, 420.9 ], "tpot_quantiles_ms": [ 52.2, 73.9, 82.8, 88.6, 90.8, 93.4, 95.1, 95.9, 97.3, 98.7, 99.7, 102.7, 103.5, 103.7, 104.3, 104.8, 106.9, 109.5, 109.9, 110.9, 111.6, 112.9, 114.3, 115.2, 115.9, 117.0, 118.6, 120.1, 120.9, 121.6, 122.4, 123.1, 123.4, 124.3, 125.6, 126.3, 127.9, 129.1, 131.0, 132.7, 133.4, 134.9, 135.7, 135.7, 136.9, 137.2, 137.5, 138.5, 139.2, 139.4, 139.6 ] }, { "x": 4.0, "output_throughput": { "median": 364.73, "min": 344.54, "max": 367.23, "n": 3 }, "median_ttft_ms": { "median": 277.72, "min": 274.53, "max": 282.89, "n": 3 }, "p99_ttft_ms": { "median": 463.95, "min": 437.73, "max": 557.07, "n": 3 }, "median_tpot_ms": { "median": 170.91, "min": 164.49, "max": 173.71, "n": 3 }, "p99_tpot_ms": { "median": 188.58, "min": 188.38, "max": 192.16, "n": 3 }, "p99_itl_ms": { "median": 621.38, "min": 452.24, "max": 762.11, "n": 3 }, "median_itl_ms": { "median": 133.15, "min": 132.0, "max": 139.05, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 2.85, "min": 2.69, "max": 2.87, "n": 3 }, "completed": [ 120, 120, 120 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 40.6, 54.6, 59.2, 68.2, 78.1, 85.6, 100.6, 108.2, 114.1, 122.2, 135.0, 145.5, 153.5, 168.6, 173.9, 178.8, 184.1, 205.7, 220.0, 225.7, 233.2, 242.7, 248.5, 254.3, 267.4, 277.6, 280.9, 288.2, 296.2, 301.5, 307.1, 310.3, 317.6, 323.1, 333.1, 341.1, 349.9, 354.2, 361.5, 370.7, 378.6, 385.2, 388.1, 391.4, 396.4, 412.7, 422.8, 432.6, 450.4, 463.9, 653.1 ], "tpot_quantiles_ms": [ 93.2, 115.3, 119.1, 122.1, 124.0, 127.2, 130.1, 132.4, 135.6, 138.1, 141.5, 144.0, 145.8, 147.1, 149.0, 151.0, 153.8, 156.0, 157.8, 161.2, 163.2, 164.2, 165.0, 166.8, 167.8, 170.4, 171.9, 172.1, 173.8, 175.5, 177.4, 179.0, 180.9, 182.4, 183.7, 184.8, 185.8, 186.3, 186.5, 186.6, 186.7, 187.2, 187.7, 188.0, 188.2, 188.4, 188.8, 191.0, 191.3, 191.6, 192.3 ] }, { "x": 8.0, "output_throughput": { "median": 639.31, "min": 611.42, "max": 640.74, "n": 3 }, "median_ttft_ms": { "median": 304.52, "min": 300.88, "max": 308.98, "n": 3 }, "p99_ttft_ms": { "median": 789.04, "min": 672.89, "max": 811.84, "n": 3 }, "median_tpot_ms": { "median": 244.32, "min": 240.2, "max": 248.7, "n": 3 }, "p99_tpot_ms": { "median": 254.83, "min": 253.99, "max": 260.06, "n": 3 }, "p99_itl_ms": { "median": 1540.95, "min": 1535.06, "max": 1807.23, "n": 3 }, "median_itl_ms": { "median": 166.9, "min": 164.68, "max": 169.47, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 4.99, "min": 4.78, "max": 5.01, "n": 3 }, "completed": [ 240, 240, 240 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 40.7, 53.6, 67.0, 75.8, 80.5, 94.0, 108.5, 117.3, 126.3, 133.5, 139.3, 153.4, 165.3, 179.9, 188.5, 194.6, 206.0, 214.2, 222.3, 233.7, 241.4, 253.0, 260.5, 278.8, 292.7, 305.2, 318.1, 330.6, 342.9, 365.3, 374.9, 390.0, 402.8, 418.1, 427.9, 436.7, 448.8, 459.6, 466.5, 477.6, 486.3, 499.6, 512.6, 522.4, 536.7, 561.0, 587.8, 610.6, 641.9, 696.1, 898.7 ], "tpot_quantiles_ms": [ 153.5, 161.8, 166.3, 170.8, 174.2, 176.6, 181.8, 185.2, 189.7, 193.7, 197.6, 201.4, 204.2, 208.8, 213.4, 216.1, 219.7, 223.1, 225.3, 228.5, 231.4, 234.3, 238.3, 240.7, 241.9, 244.5, 245.8, 246.7, 247.8, 248.5, 250.0, 250.7, 251.2, 251.9, 252.6, 252.9, 253.2, 253.3, 253.5, 253.6, 254.0, 254.2, 254.5, 254.9, 255.2, 256.8, 257.6, 258.2, 259.2, 259.8, 260.5 ] }, { "x": 16.0, "output_throughput": { "median": 1061.38, "min": 1008.87, "max": 1061.47, "n": 3 }, "median_ttft_ms": { "median": 385.1, "min": 373.53, "max": 480.84, "n": 3 }, "p99_ttft_ms": { "median": 2394.14, "min": 1671.28, "max": 2951.15, "n": 3 }, "median_tpot_ms": { "median": 338.72, "min": 337.55, "max": 343.39, "n": 3 }, "p99_tpot_ms": { "median": 352.27, "min": 350.99, "max": 360.55, "n": 3 }, "p99_itl_ms": { "median": 688.63, "min": 368.9, "max": 2226.66, "n": 3 }, "median_itl_ms": { "median": 223.37, "min": 223.02, "max": 225.19, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 8.29, "min": 7.88, "max": 8.29, "n": 3 }, "completed": [ 480, 480, 480 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2392.0, "max": 2405.0, "n": 3 }, "max_temp_c": 88.0, "thermal_share": 0.014 }, "ttft_quantiles_ms": [ 40.0, 59.2, 71.7, 79.8, 88.7, 97.7, 104.4, 113.4, 122.7, 130.6, 140.2, 155.4, 164.7, 179.1, 188.4, 201.0, 213.3, 228.2, 237.0, 248.7, 261.6, 283.4, 300.5, 327.3, 346.2, 386.9, 530.1, 607.5, 675.8, 739.5, 810.7, 865.6, 917.9, 973.9, 1031.8, 1098.4, 1199.5, 1281.8, 1363.9, 1464.9, 1572.2, 1635.1, 1748.4, 1855.7, 1949.6, 2062.5, 2165.9, 2284.1, 2440.4, 2694.5, 3091.2 ], "tpot_quantiles_ms": [ 214.7, 224.2, 228.7, 232.0, 236.6, 240.3, 245.3, 251.2, 255.5, 257.8, 265.3, 269.2, 273.3, 279.6, 283.7, 289.7, 292.8, 298.9, 305.9, 311.5, 318.7, 323.4, 328.6, 333.5, 337.6, 339.7, 341.1, 341.9, 342.8, 344.2, 345.2, 346.5, 347.4, 347.8, 348.4, 348.7, 349.2, 349.6, 350.2, 350.7, 351.4, 351.8, 352.1, 352.6, 355.5, 357.3, 358.7, 359.2, 359.8, 360.2, 360.7 ] } ] } }, "quantiles": [ 0.0, 0.02, 0.04, 0.06, 0.08, 0.1, 0.12, 0.14, 0.16, 0.18, 0.2, 0.22, 0.24, 0.26, 0.28, 0.3, 0.32, 0.34, 0.36, 0.38, 0.4, 0.42, 0.44, 0.46, 0.48, 0.5, 0.52, 0.54, 0.56, 0.58, 0.6, 0.62, 0.64, 0.66, 0.68, 0.7, 0.72, 0.74, 0.76, 0.78, 0.8, 0.82, 0.84, 0.86, 0.88, 0.9, 0.92, 0.94, 0.96, 0.98, 1.0 ], "waves": { "default": { "kv_pool_tokens": 256063, "points": [ { "x": 0.0, "output_throughput": { "median": 94.48, "min": 91.31, "max": 95.12, "n": 3 }, "median_ttft_ms": { "median": 319.91, "min": 254.58, "max": 325.49, "n": 3 }, "p99_ttft_ms": { "median": 386.42, "min": 376.68, "max": 395.09, "n": 3 }, "median_tpot_ms": { "median": 158.28, "min": 155.53, "max": 162.31, "n": 3 }, "p99_tpot_ms": { "median": 174.38, "min": 160.21, "max": 178.24, "n": 3 }, "p99_itl_ms": { "median": 229.59, "min": 204.23, "max": 231.9, "n": 3 }, "median_itl_ms": { "median": 154.73, "min": 153.04, "max": 158.57, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.05, "n": 3 }, "request_throughput": { "median": 1.33, "min": 1.29, "max": 1.57, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 49.7, 64.2, 107.5, 184.8, 191.1, 193.0, 193.2, 193.8, 200.9, 201.0, 201.2, 201.4, 205.9, 207.2, 207.4, 207.6, 207.8, 214.2, 216.5, 225.2, 231.2, 250.4, 254.5, 270.3, 315.3, 317.9, 323.6, 328.1, 332.6, 334.1, 336.2, 337.9, 338.8, 340.2, 341.1, 344.0, 347.5, 351.2, 353.3, 353.9, 355.3, 358.5, 361.2, 362.0, 364.2, 366.6, 373.2, 376.7, 378.8, 386.2, 395.6 ], "tpot_quantiles_ms": [ 91.7, 117.1, 127.5, 132.7, 140.8, 145.2, 147.2, 148.9, 151.0, 152.9, 153.3, 153.6, 153.9, 154.4, 154.9, 155.1, 155.3, 155.6, 155.8, 156.0, 156.6, 156.8, 157.0, 157.3, 157.8, 158.3, 158.5, 158.7, 159.2, 159.3, 159.4, 159.7, 160.0, 160.8, 161.4, 162.0, 162.2, 162.5, 162.7, 163.1, 164.2, 164.9, 165.2, 165.7, 166.1, 166.6, 167.4, 167.8, 168.8, 171.5, 186.4 ] }, { "x": 0.5, "output_throughput": { "median": 97.8, "min": 96.82, "max": 98.35, "n": 3 }, "median_ttft_ms": { "median": 321.79, "min": 316.01, "max": 326.49, "n": 3 }, "p99_ttft_ms": { "median": 367.95, "min": 361.18, "max": 373.77, "n": 3 }, "median_tpot_ms": { "median": 156.44, "min": 153.16, "max": 158.06, "n": 3 }, "p99_tpot_ms": { "median": 159.96, "min": 158.65, "max": 163.32, "n": 3 }, "p99_itl_ms": { "median": 200.6, "min": 195.34, "max": 206.78, "n": 3 }, "median_itl_ms": { "median": 153.2, "min": 150.55, "max": 154.79, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 1.0, "min": 0.99, "max": 1.03, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 52.0, 73.1, 193.1, 196.6, 213.7, 218.9, 219.3, 219.7, 220.4, 220.5, 220.5, 220.5, 220.7, 222.0, 222.1, 222.1, 222.2, 222.3, 227.3, 250.8, 307.3, 311.9, 313.1, 314.4, 318.8, 321.4, 324.3, 327.6, 328.7, 330.4, 331.4, 334.5, 334.9, 336.2, 338.4, 339.4, 341.3, 343.1, 344.6, 344.9, 348.6, 349.7, 352.3, 354.8, 355.6, 356.2, 356.8, 360.5, 366.4, 370.9, 374.2 ], "tpot_quantiles_ms": [ 91.2, 108.2, 119.9, 135.4, 140.8, 144.6, 144.8, 146.0, 147.1, 147.5, 147.8, 148.0, 148.9, 149.4, 149.7, 150.2, 150.5, 150.8, 151.1, 151.6, 152.6, 152.8, 153.3, 153.6, 154.1, 154.2, 154.7, 155.0, 155.2, 155.6, 155.9, 156.3, 156.6, 156.9, 157.1, 157.3, 157.4, 157.6, 157.7, 157.9, 158.4, 158.7, 159.1, 159.8, 160.5, 160.7, 160.9, 161.2, 161.5, 161.9, 164.4 ] }, { "x": 1.0, "output_throughput": { "median": 99.85, "min": 98.54, "max": 101.33, "n": 3 }, "median_ttft_ms": { "median": 1649.77, "min": 1604.56, "max": 1653.14, "n": 3 }, "p99_ttft_ms": { "median": 1715.8, "min": 1702.17, "max": 1973.58, "n": 3 }, "median_tpot_ms": { "median": 151.37, "min": 150.76, "max": 155.29, "n": 3 }, "p99_tpot_ms": { "median": 163.39, "min": 162.94, "max": 168.4, "n": 3 }, "p99_itl_ms": { "median": 169.46, "min": 166.04, "max": 172.94, "n": 3 }, "median_itl_ms": { "median": 149.3, "min": 144.91, "max": 150.16, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.78, "min": 0.77, "max": 0.79, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 54.3, 268.5, 301.3, 480.2, 509.5, 538.8, 555.9, 605.1, 642.2, 719.2, 1158.7, 1158.8, 1210.9, 1256.2, 1256.7, 1547.7, 1548.6, 1549.2, 1602.9, 1604.5, 1605.5, 1606.5, 1639.5, 1639.8, 1640.7, 1641.3, 1648.4, 1649.4, 1650.0, 1665.0, 1665.2, 1665.5, 1673.2, 1686.8, 1687.2, 1687.7, 1693.6, 1693.7, 1701.0, 1701.7, 1702.2, 1712.2, 1713.0, 1714.0, 1716.7, 1926.1, 1926.7, 1927.4, 1972.6, 1973.3, 1973.7 ], "tpot_quantiles_ms": [ 133.9, 133.9, 133.9, 133.9, 139.8, 139.8, 139.8, 142.3, 142.3, 142.3, 142.3, 142.5, 142.5, 142.5, 142.5, 146.3, 146.3, 146.3, 146.3, 147.6, 149.1, 149.1, 149.2, 150.7, 150.8, 150.8, 150.8, 150.9, 151.4, 151.4, 151.4, 152.6, 153.7, 153.8, 153.8, 155.3, 155.9, 155.9, 157.5, 157.5, 157.5, 157.5, 159.5, 159.5, 159.7, 160.3, 161.8, 162.5, 164.4, 165.8, 168.9 ] } ] }, "fp8": { "kv_pool_tokens": 1080828, "points": [ { "x": 0.0, "output_throughput": { "median": 172.25, "min": 166.07, "max": 172.98, "n": 3 }, "median_ttft_ms": { "median": 187.89, "min": 185.87, "max": 196.39, "n": 3 }, "p99_ttft_ms": { "median": 229.73, "min": 223.67, "max": 237.7, "n": 3 }, "median_tpot_ms": { "median": 87.04, "min": 84.67, "max": 89.17, "n": 3 }, "p99_tpot_ms": { "median": 93.14, "min": 87.0, "max": 100.23, "n": 3 }, "p99_itl_ms": { "median": 126.91, "min": 116.32, "max": 140.79, "n": 3 }, "median_itl_ms": { "median": 82.98, "min": 81.05, "max": 83.79, "n": 3 }, "goodput_rps": { "median": 2.42, "min": 2.32, "max": 2.85, "n": 3 }, "request_throughput": { "median": 2.42, "min": 2.35, "max": 2.85, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 51.6, 64.3, 128.3, 128.5, 128.6, 129.0, 130.5, 130.6, 130.6, 130.9, 135.3, 135.7, 136.2, 136.8, 137.5, 141.6, 143.4, 144.0, 145.0, 149.1, 153.2, 173.0, 184.1, 184.8, 186.9, 188.1, 189.2, 190.2, 191.3, 191.7, 193.3, 194.8, 195.2, 196.4, 197.1, 197.9, 198.1, 198.6, 198.8, 199.5, 200.1, 200.6, 203.3, 204.5, 206.2, 214.0, 217.2, 222.0, 225.1, 227.2, 241.5 ], "tpot_quantiles_ms": [ 50.0, 64.5, 68.3, 72.0, 75.8, 78.8, 79.3, 81.5, 82.3, 83.0, 83.3, 83.4, 83.7, 84.1, 84.2, 84.4, 84.6, 84.7, 85.0, 85.1, 85.2, 85.3, 85.5, 85.6, 85.7, 85.8, 86.1, 86.3, 86.4, 86.6, 87.0, 87.1, 87.4, 87.7, 88.1, 88.2, 88.7, 88.8, 89.1, 89.3, 89.7, 89.7, 89.8, 90.1, 90.3, 90.5, 90.7, 91.4, 92.4, 93.8, 102.7 ] }, { "x": 0.5, "output_throughput": { "median": 178.63, "min": 177.17, "max": 179.55, "n": 3 }, "median_ttft_ms": { "median": 187.73, "min": 184.97, "max": 190.76, "n": 3 }, "p99_ttft_ms": { "median": 226.93, "min": 220.72, "max": 227.92, "n": 3 }, "median_tpot_ms": { "median": 84.67, "min": 82.18, "max": 84.8, "n": 3 }, "p99_tpot_ms": { "median": 87.18, "min": 86.23, "max": 89.01, "n": 3 }, "p99_itl_ms": { "median": 115.42, "min": 114.35, "max": 119.47, "n": 3 }, "median_itl_ms": { "median": 81.33, "min": 80.15, "max": 82.51, "n": 3 }, "goodput_rps": { "median": 1.84, "min": 1.8, "max": 1.89, "n": 3 }, "request_throughput": { "median": 1.84, "min": 1.8, "max": 1.89, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 51.0, 61.8, 127.6, 127.7, 127.9, 128.0, 131.1, 131.2, 131.3, 131.6, 133.0, 133.1, 133.2, 133.5, 140.5, 142.1, 146.4, 150.9, 170.9, 173.5, 180.5, 181.9, 183.8, 184.5, 186.3, 186.8, 188.1, 189.5, 190.7, 191.4, 192.4, 192.6, 193.7, 194.4, 194.7, 195.7, 196.1, 196.7, 197.0, 198.2, 198.7, 199.1, 199.9, 200.5, 201.3, 203.1, 211.2, 216.0, 221.1, 226.1, 228.8 ], "tpot_quantiles_ms": [ 51.9, 58.4, 66.0, 74.3, 76.3, 78.3, 79.1, 79.9, 80.4, 80.8, 80.9, 81.4, 81.5, 81.6, 81.9, 82.1, 82.3, 82.3, 82.5, 82.6, 82.7, 83.4, 83.6, 83.6, 83.8, 84.1, 84.2, 84.3, 84.4, 84.5, 84.7, 84.7, 84.9, 85.0, 85.1, 85.2, 85.2, 85.3, 85.4, 85.5, 85.9, 86.0, 86.1, 86.5, 86.6, 86.6, 86.8, 87.2, 87.5, 88.0, 90.1 ] }, { "x": 1.0, "output_throughput": { "median": 183.48, "min": 181.7, "max": 187.43, "n": 3 }, "median_ttft_ms": { "median": 1050.47, "min": 1042.66, "max": 1094.67, "n": 3 }, "p99_ttft_ms": { "median": 1176.76, "min": 1151.25, "max": 1188.43, "n": 3 }, "median_tpot_ms": { "median": 80.52, "min": 80.36, "max": 82.11, "n": 3 }, "p99_tpot_ms": { "median": 88.77, "min": 88.54, "max": 90.6, "n": 3 }, "p99_itl_ms": { "median": 91.74, "min": 86.21, "max": 94.18, "n": 3 }, "median_itl_ms": { "median": 79.58, "min": 77.94, "max": 79.66, "n": 3 }, "goodput_rps": { "median": 0.72, "min": 0.67, "max": 0.73, "n": 3 }, "request_throughput": { "median": 1.43, "min": 1.42, "max": 1.46, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 85.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 52.7, 120.8, 149.9, 162.7, 358.4, 390.8, 399.1, 438.4, 463.0, 630.9, 631.2, 780.3, 780.5, 965.7, 965.8, 966.0, 966.2, 966.4, 966.6, 966.7, 966.9, 985.7, 985.9, 985.9, 986.0, 1094.6, 1095.3, 1095.9, 1099.7, 1100.5, 1101.4, 1134.1, 1134.5, 1135.0, 1140.7, 1140.9, 1141.5, 1150.8, 1151.0, 1151.3, 1152.3, 1152.7, 1153.3, 1174.4, 1175.5, 1176.0, 1176.3, 1176.6, 1176.8, 1188.4, 1188.5 ], "tpot_quantiles_ms": [ 71.2, 71.2, 74.4, 74.4, 74.7, 74.7, 74.7, 74.7, 75.8, 75.8, 75.8, 76.2, 76.2, 76.2, 78.3, 78.3, 79.3, 79.8, 79.8, 79.8, 79.8, 80.4, 80.4, 80.5, 80.5, 80.5, 80.5, 81.2, 81.4, 81.4, 81.5, 81.5, 81.6, 82.0, 82.0, 82.0, 82.1, 82.1, 82.1, 82.1, 83.5, 83.7, 83.7, 83.7, 84.1, 84.5, 84.5, 87.6, 88.2, 89.8, 91.8 ] } ] }, "mixed-chunk": { "kv_pool_tokens": 259786, "points": [ { "x": 0.0, "output_throughput": { "median": 98.11, "min": 95.01, "max": 99.64, "n": 3 }, "median_ttft_ms": { "median": 441.96, "min": 424.72, "max": 443.02, "n": 3 }, "p99_ttft_ms": { "median": 579.63, "min": 575.93, "max": 592.33, "n": 3 }, "median_tpot_ms": { "median": 148.47, "min": 144.76, "max": 150.28, "n": 3 }, "p99_tpot_ms": { "median": 170.45, "min": 156.38, "max": 174.59, "n": 3 }, "p99_itl_ms": { "median": 201.65, "min": 194.22, "max": 201.81, "n": 3 }, "median_itl_ms": { "median": 147.73, "min": 144.28, "max": 149.75, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.0, "max": 0.05, "n": 3 }, "request_throughput": { "median": 1.4, "min": 1.33, "max": 1.63, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 79.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 50.2, 194.9, 195.0, 195.0, 195.2, 199.1, 199.5, 200.1, 200.8, 221.3, 222.1, 222.6, 223.2, 407.4, 412.8, 416.6, 418.8, 419.7, 421.9, 424.7, 425.7, 428.3, 431.6, 433.9, 435.0, 435.8, 436.3, 437.4, 438.1, 438.9, 440.6, 441.9, 443.6, 445.7, 447.5, 448.6, 449.2, 451.8, 452.9, 453.9, 454.6, 456.7, 460.1, 460.8, 462.0, 464.4, 469.7, 479.8, 572.1, 581.3, 594.0 ], "tpot_quantiles_ms": [ 92.3, 113.5, 123.1, 131.2, 137.1, 140.0, 140.7, 141.1, 141.7, 142.3, 142.5, 143.0, 143.6, 144.2, 144.5, 144.9, 145.0, 145.5, 145.9, 146.3, 146.5, 146.5, 146.8, 147.4, 147.7, 147.8, 148.2, 148.4, 148.6, 148.8, 149.7, 150.2, 150.4, 150.8, 151.0, 151.6, 152.2, 152.4, 152.8, 153.6, 154.0, 154.6, 154.9, 155.3, 155.6, 156.6, 157.2, 158.8, 160.3, 162.4, 200.3 ] }, { "x": 0.5, "output_throughput": { "median": 102.27, "min": 102.21, "max": 103.88, "n": 3 }, "median_ttft_ms": { "median": 431.52, "min": 418.51, "max": 432.93, "n": 3 }, "p99_ttft_ms": { "median": 468.0, "min": 467.61, "max": 470.3, "n": 3 }, "median_tpot_ms": { "median": 146.68, "min": 143.97, "max": 148.76, "n": 3 }, "p99_tpot_ms": { "median": 150.83, "min": 147.9, "max": 156.96, "n": 3 }, "p99_itl_ms": { "median": 164.41, "min": 161.0, "max": 202.86, "n": 3 }, "median_itl_ms": { "median": 146.04, "min": 142.2, "max": 146.6, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.03, "n": 3 }, "request_throughput": { "median": 1.06, "min": 1.03, "max": 1.09, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 49.5, 199.8, 200.0, 200.1, 200.2, 203.2, 203.3, 203.5, 203.6, 210.3, 210.8, 211.4, 212.1, 386.6, 393.0, 396.5, 403.4, 406.1, 408.3, 413.1, 415.1, 416.2, 418.4, 420.8, 425.8, 427.3, 429.6, 433.0, 433.7, 434.2, 435.3, 437.0, 438.1, 440.4, 440.6, 441.8, 442.8, 443.5, 444.6, 445.7, 446.6, 447.8, 449.3, 451.7, 452.5, 455.4, 456.7, 458.8, 462.8, 467.8, 470.7 ], "tpot_quantiles_ms": [ 87.4, 103.3, 113.9, 129.1, 133.1, 134.7, 135.3, 135.9, 136.9, 137.5, 138.0, 138.6, 139.1, 139.5, 139.7, 140.1, 140.2, 140.5, 140.9, 141.3, 141.5, 143.0, 143.7, 144.0, 144.8, 144.9, 145.2, 146.2, 146.3, 146.4, 146.6, 146.7, 146.9, 147.1, 147.3, 147.4, 147.7, 147.9, 148.5, 148.7, 149.3, 149.9, 150.4, 150.7, 150.8, 151.1, 151.6, 155.4, 156.3, 156.7, 157.1 ] }, { "x": 1.0, "output_throughput": { "median": 101.62, "min": 101.24, "max": 102.59, "n": 3 }, "median_ttft_ms": { "median": 1515.59, "min": 1293.8, "max": 1596.96, "n": 3 }, "p99_ttft_ms": { "median": 1845.25, "min": 1745.38, "max": 2135.57, "n": 3 }, "median_tpot_ms": { "median": 149.17, "min": 148.06, "max": 151.15, "n": 3 }, "p99_tpot_ms": { "median": 159.77, "min": 156.98, "max": 160.38, "n": 3 }, "p99_itl_ms": { "median": 209.63, "min": 171.79, "max": 294.99, "n": 3 }, "median_itl_ms": { "median": 145.01, "min": 142.07, "max": 145.52, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.79, "min": 0.79, "max": 0.8, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.011 }, "ttft_quantiles_ms": [ 63.5, 468.3, 476.3, 568.1, 598.6, 708.1, 730.4, 808.8, 831.7, 834.8, 874.7, 921.0, 1007.6, 1007.7, 1007.9, 1069.6, 1124.3, 1133.2, 1272.8, 1273.9, 1274.8, 1292.7, 1293.9, 1294.6, 1298.3, 1298.8, 1299.3, 1492.2, 1514.9, 1515.3, 1515.9, 1517.0, 1538.1, 1593.4, 1594.3, 1596.2, 1597.0, 1598.2, 1696.5, 1697.0, 1697.4, 1698.1, 1699.4, 1723.8, 1732.6, 1745.2, 1745.4, 1844.6, 2035.6, 2035.7, 2305.3 ], "tpot_quantiles_ms": [ 134.1, 134.1, 137.4, 137.4, 137.4, 138.0, 138.0, 138.7, 138.8, 138.8, 138.9, 138.9, 142.7, 142.9, 142.9, 143.5, 144.1, 144.9, 145.7, 147.6, 147.7, 148.6, 148.6, 148.8, 149.2, 149.2, 149.7, 149.7, 149.7, 150.5, 150.5, 150.9, 151.2, 151.2, 151.2, 152.9, 153.0, 153.0, 153.0, 153.0, 153.7, 154.2, 154.2, 154.2, 154.6, 155.7, 156.1, 156.7, 157.6, 159.6, 161.6 ] } ] }, "graphs-on-b": { "kv_pool_tokens": 234958, "points": [ { "x": 0.0, "output_throughput": { "median": 96.6, "min": 93.38, "max": 96.96, "n": 3 }, "median_ttft_ms": { "median": 324.4, "min": 318.91, "max": 324.79, "n": 3 }, "p99_ttft_ms": { "median": 383.93, "min": 381.27, "max": 404.63, "n": 3 }, "median_tpot_ms": { "median": 155.55, "min": 152.43, "max": 158.58, "n": 3 }, "p99_tpot_ms": { "median": 166.69, "min": 157.2, "max": 170.89, "n": 3 }, "p99_itl_ms": { "median": 213.67, "min": 190.63, "max": 224.61, "n": 3 }, "median_itl_ms": { "median": 151.53, "min": 149.52, "max": 154.82, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.05, "n": 3 }, "request_throughput": { "median": 1.36, "min": 1.32, "max": 1.6, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 49.1, 71.3, 181.8, 195.4, 195.5, 195.6, 195.7, 196.6, 200.6, 200.7, 200.7, 200.9, 204.0, 208.0, 210.8, 211.1, 211.3, 211.4, 214.0, 222.3, 237.2, 254.3, 267.8, 308.4, 316.5, 320.7, 325.0, 327.9, 329.3, 330.0, 330.6, 332.2, 336.0, 337.8, 339.3, 342.0, 343.4, 344.6, 345.8, 346.9, 348.2, 349.9, 351.3, 355.3, 356.8, 357.9, 362.0, 369.0, 374.5, 382.0, 404.9 ], "tpot_quantiles_ms": [ 89.2, 113.2, 125.1, 129.5, 137.2, 142.4, 144.1, 146.1, 148.4, 149.1, 150.4, 150.7, 151.0, 151.5, 151.9, 152.2, 152.3, 152.6, 152.8, 153.5, 153.8, 154.0, 154.1, 154.3, 154.8, 155.0, 155.4, 155.5, 155.8, 156.3, 156.4, 156.8, 157.3, 157.7, 158.1, 158.3, 158.6, 159.0, 159.7, 160.0, 161.1, 161.3, 161.4, 161.6, 162.2, 162.5, 162.6, 163.0, 164.7, 167.5, 173.9 ] }, { "x": 0.5, "output_throughput": { "median": 99.46, "min": 98.52, "max": 100.1, "n": 3 }, "median_ttft_ms": { "median": 314.84, "min": 312.37, "max": 319.1, "n": 3 }, "p99_ttft_ms": { "median": 363.39, "min": 362.36, "max": 369.46, "n": 3 }, "median_tpot_ms": { "median": 153.37, "min": 150.17, "max": 155.21, "n": 3 }, "p99_tpot_ms": { "median": 157.5, "min": 156.07, "max": 160.3, "n": 3 }, "p99_itl_ms": { "median": 197.57, "min": 192.51, "max": 201.16, "n": 3 }, "median_itl_ms": { "median": 150.85, "min": 147.57, "max": 151.97, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 1.02, "min": 1.0, "max": 1.05, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 51.5, 73.1, 181.6, 193.5, 195.2, 195.3, 195.3, 195.4, 205.0, 205.4, 205.9, 206.8, 207.2, 207.7, 207.8, 207.9, 208.6, 214.4, 218.7, 225.0, 232.5, 296.5, 304.7, 310.3, 314.0, 315.5, 319.4, 322.1, 323.2, 323.9, 325.3, 326.4, 328.1, 331.8, 333.2, 333.7, 334.9, 336.3, 337.9, 338.5, 340.0, 343.2, 344.9, 347.6, 351.0, 352.8, 354.1, 357.9, 362.7, 367.5, 369.8 ], "tpot_quantiles_ms": [ 90.0, 105.7, 118.2, 133.5, 138.2, 141.4, 142.0, 143.4, 144.0, 144.5, 145.0, 145.8, 146.1, 146.7, 147.3, 147.7, 148.0, 148.1, 148.6, 149.0, 149.2, 149.3, 150.3, 150.7, 151.1, 151.3, 151.7, 152.3, 152.9, 153.3, 153.4, 153.5, 153.9, 154.0, 154.1, 154.2, 154.7, 154.9, 155.1, 155.4, 155.8, 156.0, 156.2, 157.0, 157.6, 157.9, 157.9, 158.1, 158.4, 158.6, 161.5 ] }, { "x": 1.0, "output_throughput": { "median": 101.44, "min": 100.28, "max": 104.19, "n": 3 }, "median_ttft_ms": { "median": 1275.38, "min": 1268.0, "max": 1626.6, "n": 3 }, "p99_ttft_ms": { "median": 1722.69, "min": 1719.0, "max": 1729.81, "n": 3 }, "median_tpot_ms": { "median": 149.7, "min": 147.69, "max": 152.29, "n": 3 }, "p99_tpot_ms": { "median": 160.33, "min": 159.35, "max": 164.69, "n": 3 }, "p99_itl_ms": { "median": 159.48, "min": 159.32, "max": 189.5, "n": 3 }, "median_itl_ms": { "median": 147.92, "min": 143.69, "max": 148.15, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.79, "min": 0.78, "max": 0.81, "n": 3 }, "completed": [ 64, 64, 64 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.011 }, "ttft_quantiles_ms": [ 50.8, 206.7, 268.9, 291.9, 613.0, 731.9, 818.5, 818.6, 840.9, 850.2, 900.1, 900.7, 1042.5, 1042.8, 1267.6, 1267.8, 1267.9, 1268.0, 1271.4, 1272.0, 1272.5, 1273.3, 1275.1, 1275.1, 1275.3, 1275.5, 1626.4, 1626.9, 1627.3, 1650.9, 1651.0, 1651.2, 1688.1, 1688.3, 1689.0, 1689.8, 1690.8, 1706.4, 1707.1, 1707.8, 1708.2, 1714.9, 1715.0, 1717.5, 1718.7, 1719.0, 1722.5, 1722.7, 1728.2, 1729.0, 1729.9 ], "tpot_quantiles_ms": [ 131.3, 131.3, 131.3, 137.3, 137.3, 138.2, 139.4, 139.4, 139.4, 139.4, 140.1, 140.1, 142.6, 142.6, 142.7, 144.7, 144.7, 144.7, 146.4, 147.6, 147.6, 147.6, 147.7, 147.7, 147.7, 147.7, 148.6, 149.7, 149.7, 149.7, 149.7, 150.6, 150.6, 150.7, 150.7, 150.7, 150.7, 152.3, 152.3, 152.3, 152.3, 152.6, 156.1, 156.1, 156.1, 156.2, 157.1, 158.9, 161.9, 162.9, 167.7 ] } ] } }, "working-set": { "default": { "kv_pool_tokens": 260013, "points": [ { "x": 4.0, "output_throughput": { "median": 104.9, "min": 97.77, "max": 110.8, "n": 3 }, "median_ttft_ms": { "median": 456.69, "min": 445.79, "max": 495.32, "n": 3 }, "p99_ttft_ms": { "median": 2088.17, "min": 1716.6, "max": 2145.88, "n": 3 }, "median_tpot_ms": { "median": 53.86, "min": 53.54, "max": 57.58, "n": 3 }, "p99_tpot_ms": { "median": 98.54, "min": 89.88, "max": 106.3, "n": 3 }, "p99_itl_ms": { "median": 69.03, "min": 68.22, "max": 420.42, "n": 3 }, "median_itl_ms": { "median": 49.92, "min": 45.47, "max": 50.66, "n": 3 }, "goodput_rps": { "median": 2.58, "min": 2.46, "max": 2.92, "n": 3 }, "request_throughput": { "median": 3.28, "min": 3.06, "max": 3.46, "n": 3 }, "completed": [ 32, 32, 32 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2405.0, "max": 2411.0, "n": 3 }, "max_temp_c": 82.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 54.9, 83.3, 174.2, 211.7, 220.3, 259.5, 266.1, 274.2, 277.8, 438.3, 438.6, 438.8, 439.2, 444.7, 444.9, 445.6, 446.7, 452.8, 453.0, 453.2, 454.9, 455.0, 455.4, 455.7, 456.2, 456.3, 456.6, 457.3, 467.3, 467.9, 468.6, 469.1, 488.5, 489.3, 494.6, 495.5, 505.8, 506.6, 611.9, 629.3, 684.5, 789.1, 1175.4, 1277.7, 1343.0, 1716.3, 1716.6, 2088.1, 2088.2, 2145.8, 2145.9 ], "tpot_quantiles_ms": [ 42.4, 42.4, 45.3, 45.3, 45.3, 45.3, 46.1, 46.1, 46.1, 46.4, 50.4, 50.4, 51.3, 51.4, 51.4, 51.4, 52.8, 52.8, 53.1, 53.5, 53.5, 53.7, 53.7, 53.8, 53.8, 53.9, 53.9, 53.9, 54.4, 55.1, 55.3, 55.3, 55.3, 55.3, 57.3, 57.4, 57.8, 59.0, 59.0, 59.9, 61.2, 61.4, 63.1, 64.5, 66.3, 77.6, 77.6, 78.0, 95.2, 96.6, 119.2 ] }, { "x": 16.0, "output_throughput": { "median": 106.38, "min": 98.95, "max": 106.62, "n": 3 }, "median_ttft_ms": { "median": 459.98, "min": 452.16, "max": 466.54, "n": 3 }, "p99_ttft_ms": { "median": 1828.68, "min": 1806.88, "max": 4245.72, "n": 3 }, "median_tpot_ms": { "median": 55.36, "min": 55.23, "max": 55.68, "n": 3 }, "p99_tpot_ms": { "median": 100.49, "min": 92.31, "max": 163.66, "n": 3 }, "p99_itl_ms": { "median": 74.2, "min": 73.48, "max": 77.69, "n": 3 }, "median_itl_ms": { "median": 53.71, "min": 52.51, "max": 54.1, "n": 3 }, "goodput_rps": { "median": 2.57, "min": 2.54, "max": 2.76, "n": 3 }, "request_throughput": { "median": 3.32, "min": 3.09, "max": 3.33, "n": 3 }, "completed": [ 128, 128, 128 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 83.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 65.6, 197.6, 202.8, 207.8, 214.1, 222.8, 253.3, 438.2, 442.9, 443.8, 444.6, 445.1, 446.1, 447.0, 449.3, 450.7, 452.5, 453.9, 455.0, 455.7, 456.2, 457.0, 457.8, 458.7, 459.2, 459.7, 460.5, 461.9, 462.4, 463.6, 466.2, 466.8, 467.4, 470.8, 472.2, 478.6, 509.4, 599.4, 776.7, 799.9, 822.1, 974.5, 1121.1, 1125.4, 1430.1, 1437.5, 1480.1, 1577.0, 1810.1, 2307.2, 4246.0 ], "tpot_quantiles_ms": [ 45.2, 45.6, 46.6, 46.8, 46.8, 47.3, 47.6, 47.8, 48.0, 48.1, 48.4, 48.7, 48.8, 49.1, 49.3, 49.8, 49.9, 50.7, 52.5, 52.9, 53.7, 54.0, 54.7, 54.9, 55.3, 55.7, 55.7, 55.8, 56.0, 56.4, 56.8, 57.0, 57.2, 57.3, 57.8, 59.1, 60.9, 61.0, 61.5, 62.6, 63.1, 63.7, 64.0, 64.6, 64.7, 64.9, 68.7, 72.2, 77.1, 94.3, 184.4 ] }, { "x": 32.0, "output_throughput": { "median": 110.0, "min": 108.96, "max": 111.88, "n": 3 }, "median_ttft_ms": { "median": 450.93, "min": 443.76, "max": 455.64, "n": 3 }, "p99_ttft_ms": { "median": 1982.35, "min": 1808.14, "max": 2312.29, "n": 3 }, "median_tpot_ms": { "median": 55.08, "min": 52.47, "max": 55.62, "n": 3 }, "p99_tpot_ms": { "median": 86.0, "min": 85.12, "max": 107.63, "n": 3 }, "p99_itl_ms": { "median": 78.71, "min": 76.16, "max": 78.93, "n": 3 }, "median_itl_ms": { "median": 53.84, "min": 51.65, "max": 54.59, "n": 3 }, "goodput_rps": { "median": 3.03, "min": 2.93, "max": 3.25, "n": 3 }, "request_throughput": { "median": 3.44, "min": 3.41, "max": 3.5, "n": 3 }, "completed": [ 256, 256, 256 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 84.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 56.3, 196.3, 200.9, 205.3, 210.0, 218.5, 432.5, 435.6, 437.9, 438.7, 439.7, 440.3, 441.5, 442.8, 443.5, 444.1, 444.5, 444.9, 445.5, 446.4, 447.5, 448.3, 449.2, 449.8, 450.4, 451.1, 452.6, 453.7, 454.3, 454.7, 455.2, 455.8, 457.0, 459.7, 460.7, 461.5, 463.6, 465.2, 505.6, 613.9, 792.7, 795.3, 803.4, 804.6, 806.4, 1074.8, 1122.8, 1151.2, 1435.3, 1781.4, 2389.6 ], "tpot_quantiles_ms": [ 44.9, 46.8, 47.2, 47.5, 47.6, 47.7, 47.8, 48.0, 48.0, 48.2, 48.8, 49.2, 49.7, 49.9, 50.1, 50.2, 50.4, 50.9, 51.6, 51.9, 52.4, 52.7, 53.1, 53.8, 54.5, 54.8, 55.5, 56.0, 56.2, 57.1, 57.3, 57.8, 57.9, 58.3, 58.7, 59.1, 59.4, 60.1, 61.4, 61.7, 62.2, 62.6, 63.6, 64.0, 65.6, 66.4, 68.8, 70.2, 71.9, 83.5, 131.7 ] }, { "x": 64.0, "output_throughput": { "median": 110.07, "min": 107.62, "max": 110.09, "n": 3 }, "median_ttft_ms": { "median": 448.31, "min": 448.04, "max": 460.78, "n": 3 }, "p99_ttft_ms": { "median": 1801.05, "min": 1590.3, "max": 2130.48, "n": 3 }, "median_tpot_ms": { "median": 55.9, "min": 55.25, "max": 55.94, "n": 3 }, "p99_tpot_ms": { "median": 89.11, "min": 83.63, "max": 92.52, "n": 3 }, "p99_itl_ms": { "median": 78.09, "min": 77.81, "max": 81.23, "n": 3 }, "median_itl_ms": { "median": 54.18, "min": 53.56, "max": 54.39, "n": 3 }, "goodput_rps": { "median": 3.01, "min": 2.94, "max": 3.03, "n": 3 }, "request_throughput": { "median": 3.44, "min": 3.36, "max": 3.44, "n": 3 }, "completed": [ 512, 512, 512 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.006 }, "ttft_quantiles_ms": [ 55.3, 199.8, 203.3, 207.1, 212.0, 218.5, 431.0, 435.9, 439.0, 440.2, 441.7, 442.5, 443.7, 444.5, 445.1, 445.5, 446.0, 446.4, 447.0, 447.7, 448.3, 448.9, 449.4, 450.1, 450.8, 451.4, 451.8, 453.3, 454.4, 455.9, 457.1, 458.1, 459.1, 460.0, 461.6, 464.0, 465.8, 467.7, 511.7, 777.4, 789.5, 795.2, 807.7, 815.3, 980.5, 1124.8, 1133.4, 1173.3, 1473.8, 1589.9, 2437.7 ], "tpot_quantiles_ms": [ 45.6, 47.1, 47.6, 48.3, 48.5, 48.9, 49.2, 49.5, 49.6, 50.1, 50.3, 50.5, 50.8, 50.9, 51.2, 51.3, 51.7, 51.8, 52.1, 52.4, 52.8, 53.6, 54.3, 55.1, 55.4, 55.7, 56.1, 56.5, 56.7, 57.1, 57.4, 57.9, 58.1, 58.7, 59.2, 59.5, 59.9, 60.4, 61.1, 61.8, 62.8, 63.6, 64.1, 64.7, 66.2, 68.1, 68.3, 70.1, 71.9, 81.6, 119.4 ] } ] } }, "mixed": { "chunk-1024": { "kv_pool_tokens": 241348, "points": [ { "x": 0.5, "output_throughput": { "median": 68.07, "min": 56.67, "max": 68.71, "n": 3 }, "median_ttft_ms": { "median": 1252.59, "min": 1082.72, "max": 1307.29, "n": 3 }, "p99_ttft_ms": { "median": 2647.43, "min": 2639.52, "max": 2738.34, "n": 3 }, "median_tpot_ms": { "median": 230.98, "min": 164.16, "max": 280.18, "n": 3 }, "p99_tpot_ms": { "median": 267.69, "min": 208.78, "max": 316.81, "n": 3 }, "p99_itl_ms": { "median": 1976.87, "min": 1367.03, "max": 2445.28, "n": 3 }, "median_itl_ms": { "median": 149.92, "min": 118.75, "max": 175.71, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.01, "n": 3 }, "request_throughput": { "median": 0.4, "min": 0.38, "max": 0.43, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2398.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 89.0, "thermal_share": 0.063 }, "ttft_quantiles_ms": [ 45.0, 52.9, 612.7, 667.0, 719.7, 748.3, 768.3, 802.1, 823.4, 846.7, 867.8, 880.2, 896.1, 930.8, 983.5, 994.6, 1018.6, 1040.1, 1052.5, 1059.0, 1076.9, 1095.3, 1112.9, 1129.1, 1147.7, 1224.0, 1243.5, 1285.2, 1309.2, 1322.6, 1338.3, 1359.2, 1385.3, 1406.2, 1446.4, 1478.3, 1512.6, 1552.6, 1558.1, 1595.4, 1623.6, 1635.2, 1668.2, 1735.6, 1924.5, 2064.2, 2128.3, 2269.9, 2388.8, 2571.9, 2885.9 ], "tpot_quantiles_ms": [ 87.5, 104.6, 111.7, 117.1, 125.9, 128.9, 139.5, 143.0, 148.8, 153.1, 154.7, 159.3, 163.1, 163.9, 164.9, 166.7, 170.6, 178.4, 180.4, 185.5, 188.7, 189.6, 190.8, 198.2, 201.5, 205.2, 212.3, 222.9, 229.6, 229.6, 231.7, 235.5, 240.0, 242.0, 245.3, 247.4, 250.2, 251.8, 254.3, 260.7, 264.4, 275.8, 280.4, 284.6, 292.1, 294.3, 296.9, 302.0, 303.9, 310.0, 321.8 ] }, { "x": 1.0, "output_throughput": { "median": 107.3, "min": 94.23, "max": 112.61, "n": 3 }, "median_ttft_ms": { "median": 253.05, "min": 206.72, "max": 254.48, "n": 3 }, "p99_ttft_ms": { "median": 416.9, "min": 336.54, "max": 427.81, "n": 3 }, "median_tpot_ms": { "median": 164.08, "min": 130.11, "max": 184.8, "n": 3 }, "p99_tpot_ms": { "median": 185.35, "min": 153.37, "max": 205.44, "n": 3 }, "p99_itl_ms": { "median": 228.21, "min": 175.62, "max": 243.09, "n": 3 }, "median_itl_ms": { "median": 166.08, "min": 132.23, "max": 193.79, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.0, "max": 0.08, "n": 3 }, "request_throughput": { "median": 0.66, "min": 0.63, "max": 0.67, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 44.3, 45.9, 81.1, 90.0, 102.1, 115.9, 130.1, 153.5, 161.4, 170.2, 176.0, 178.7, 184.1, 186.3, 189.4, 192.2, 197.5, 202.3, 206.1, 208.3, 209.8, 214.1, 217.1, 223.6, 228.3, 230.0, 237.3, 249.4, 250.8, 253.0, 255.1, 260.2, 266.4, 271.5, 275.8, 281.5, 288.9, 291.9, 300.6, 307.2, 311.8, 324.8, 334.6, 335.5, 349.2, 364.9, 372.7, 378.3, 387.8, 404.3, 448.6 ], "tpot_quantiles_ms": [ 67.6, 89.0, 96.9, 109.9, 113.8, 117.5, 124.7, 128.8, 129.0, 129.7, 130.2, 131.6, 135.3, 137.0, 138.1, 138.3, 139.2, 139.9, 140.8, 143.6, 144.9, 147.6, 150.2, 152.3, 153.5, 157.0, 160.1, 161.6, 163.6, 164.8, 167.3, 168.4, 170.4, 172.8, 175.3, 176.4, 176.7, 177.7, 178.5, 179.9, 181.2, 185.2, 186.1, 187.2, 190.8, 195.8, 197.0, 200.2, 202.7, 204.4, 205.6 ] }, { "x": 2.0, "output_throughput": { "median": 130.09, "min": 122.15, "max": 137.31, "n": 3 }, "median_ttft_ms": { "median": 255.86, "min": 242.04, "max": 273.19, "n": 3 }, "p99_ttft_ms": { "median": 457.44, "min": 434.38, "max": 503.66, "n": 3 }, "median_tpot_ms": { "median": 194.41, "min": 177.21, "max": 212.3, "n": 3 }, "p99_tpot_ms": { "median": 218.95, "min": 193.95, "max": 235.21, "n": 3 }, "p99_itl_ms": { "median": 260.12, "min": 227.25, "max": 279.42, "n": 3 }, "median_itl_ms": { "median": 202.19, "min": 175.32, "max": 223.6, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.82, "min": 0.81, "max": 0.82, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 46.0, 47.4, 71.6, 91.1, 100.4, 112.3, 122.5, 130.2, 134.5, 136.3, 147.1, 166.5, 185.7, 190.8, 199.7, 201.7, 202.4, 204.9, 215.9, 220.6, 232.1, 236.2, 236.7, 240.1, 251.7, 255.8, 261.9, 264.7, 266.1, 268.4, 278.4, 290.7, 295.5, 299.0, 303.4, 321.4, 331.7, 338.9, 345.1, 348.4, 351.7, 360.5, 366.8, 371.5, 373.6, 378.0, 388.4, 399.4, 414.0, 452.5, 505.8 ], "tpot_quantiles_ms": [ 106.1, 129.6, 138.5, 144.2, 147.9, 157.7, 162.8, 165.8, 166.3, 169.6, 171.6, 175.2, 176.4, 177.4, 178.5, 179.3, 180.5, 181.9, 182.7, 183.8, 184.6, 185.1, 186.4, 189.5, 189.9, 191.2, 191.6, 192.3, 193.9, 194.6, 196.8, 197.3, 198.7, 202.9, 206.1, 207.4, 207.6, 208.9, 210.2, 211.2, 212.5, 214.7, 216.3, 217.2, 219.0, 220.5, 225.4, 227.8, 229.3, 230.3, 237.7 ] } ] }, "chunk-8192": { "kv_pool_tokens": 238858, "points": [ { "x": 0.5, "output_throughput": { "median": 70.6, "min": 57.4, "max": 70.65, "n": 3 }, "median_ttft_ms": { "median": 867.67, "min": 772.52, "max": 905.02, "n": 3 }, "p99_ttft_ms": { "median": 1737.02, "min": 1610.41, "max": 1896.08, "n": 3 }, "median_tpot_ms": { "median": 175.08, "min": 129.46, "max": 213.48, "n": 3 }, "p99_tpot_ms": { "median": 204.34, "min": 176.57, "max": 252.2, "n": 3 }, "p99_itl_ms": { "median": 903.11, "min": 809.64, "max": 917.77, "n": 3 }, "median_itl_ms": { "median": 133.01, "min": 103.79, "max": 163.01, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.01, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.42, "min": 0.39, "max": 0.44, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 47.2, 87.7, 483.6, 512.8, 532.6, 544.7, 566.7, 578.4, 592.3, 607.6, 641.1, 653.5, 673.2, 689.7, 704.3, 721.9, 733.1, 741.3, 761.4, 774.7, 795.4, 807.0, 819.4, 824.6, 842.0, 856.4, 864.3, 876.2, 890.3, 896.3, 902.4, 910.5, 918.1, 929.8, 952.2, 966.6, 973.5, 994.2, 1006.1, 1048.9, 1054.8, 1126.5, 1156.7, 1178.8, 1262.9, 1309.8, 1430.1, 1522.6, 1605.0, 1697.5, 1920.9 ], "tpot_quantiles_ms": [ 74.6, 85.7, 104.7, 105.7, 108.7, 110.6, 115.4, 116.3, 117.5, 120.2, 127.3, 128.9, 131.3, 133.8, 138.5, 140.0, 142.6, 145.4, 148.3, 151.2, 153.1, 157.8, 161.2, 162.0, 163.8, 169.4, 171.3, 172.2, 174.7, 175.2, 176.1, 179.4, 181.0, 186.2, 191.1, 191.7, 192.7, 193.3, 194.5, 200.5, 207.7, 208.7, 216.5, 220.7, 225.3, 228.2, 230.2, 232.3, 237.8, 243.1, 254.1 ] }, { "x": 1.0, "output_throughput": { "median": 106.72, "min": 94.05, "max": 111.8, "n": 3 }, "median_ttft_ms": { "median": 271.79, "min": 219.93, "max": 298.96, "n": 3 }, "p99_ttft_ms": { "median": 420.12, "min": 357.47, "max": 446.12, "n": 3 }, "median_tpot_ms": { "median": 165.32, "min": 133.11, "max": 187.34, "n": 3 }, "p99_tpot_ms": { "median": 190.43, "min": 156.33, "max": 208.36, "n": 3 }, "p99_itl_ms": { "median": 229.93, "min": 183.7, "max": 243.81, "n": 3 }, "median_itl_ms": { "median": 167.11, "min": 134.44, "max": 196.34, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.0, "max": 0.08, "n": 3 }, "request_throughput": { "median": 0.66, "min": 0.63, "max": 0.67, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 46.0, 48.1, 86.0, 91.6, 106.4, 121.7, 131.7, 149.3, 158.3, 160.1, 171.9, 178.5, 181.3, 192.4, 196.0, 199.5, 207.4, 215.7, 218.1, 225.9, 229.5, 234.4, 239.4, 242.1, 246.1, 248.5, 252.7, 264.6, 269.9, 274.7, 278.2, 286.8, 296.2, 301.7, 307.8, 310.7, 313.5, 315.6, 320.9, 325.3, 340.2, 343.0, 352.2, 356.6, 365.1, 367.9, 381.8, 389.7, 410.8, 428.4, 446.6 ], "tpot_quantiles_ms": [ 67.6, 89.5, 97.5, 110.2, 114.4, 119.4, 126.1, 129.7, 130.4, 131.3, 133.5, 135.6, 137.6, 139.1, 140.2, 140.5, 141.4, 142.5, 143.7, 146.1, 146.5, 149.1, 149.8, 153.0, 156.4, 158.2, 162.1, 162.7, 164.9, 166.9, 169.7, 170.5, 172.8, 174.9, 178.1, 178.9, 179.1, 180.5, 181.5, 183.0, 184.2, 188.1, 189.8, 190.9, 193.7, 198.0, 199.0, 203.0, 205.6, 207.8, 208.4 ] }, { "x": 2.0, "output_throughput": { "median": 130.44, "min": 121.61, "max": 136.91, "n": 3 }, "median_ttft_ms": { "median": 269.48, "min": 269.38, "max": 291.86, "n": 3 }, "p99_ttft_ms": { "median": 450.52, "min": 393.71, "max": 458.67, "n": 3 }, "median_tpot_ms": { "median": 195.41, "min": 178.45, "max": 211.88, "n": 3 }, "p99_tpot_ms": { "median": 219.38, "min": 195.47, "max": 234.07, "n": 3 }, "p99_itl_ms": { "median": 280.45, "min": 233.13, "max": 288.81, "n": 3 }, "median_itl_ms": { "median": 204.43, "min": 176.29, "max": 220.85, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.82, "min": 0.8, "max": 0.82, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 45.1, 48.8, 80.1, 99.5, 113.2, 122.7, 129.8, 132.0, 142.5, 151.7, 165.4, 172.2, 179.6, 188.8, 206.1, 208.8, 217.7, 224.8, 236.0, 245.2, 254.0, 256.5, 257.6, 264.0, 268.7, 271.6, 277.6, 283.2, 285.5, 286.9, 291.7, 296.7, 306.2, 311.1, 317.8, 322.5, 326.9, 330.4, 334.4, 336.1, 347.1, 354.2, 365.7, 366.5, 372.8, 376.7, 405.8, 414.6, 439.9, 446.0, 461.6 ], "tpot_quantiles_ms": [ 106.7, 130.0, 139.3, 144.6, 149.5, 158.4, 164.4, 166.4, 167.1, 170.6, 172.1, 177.5, 178.4, 179.5, 180.3, 180.7, 182.0, 183.4, 184.5, 184.7, 185.5, 186.1, 186.8, 189.9, 190.6, 191.5, 193.4, 194.2, 194.5, 195.8, 196.3, 198.4, 201.6, 205.5, 206.8, 209.1, 209.3, 209.7, 210.8, 211.4, 212.0, 215.0, 215.8, 217.7, 218.7, 220.2, 225.1, 227.0, 228.9, 229.8, 236.4 ] } ] }, "default": { "kv_pool_tokens": 236185, "points": [ { "x": 0.5, "output_throughput": { "median": 70.05, "min": 57.42, "max": 70.27, "n": 3 }, "median_ttft_ms": { "median": 892.17, "min": 781.13, "max": 925.14, "n": 3 }, "p99_ttft_ms": { "median": 1753.19, "min": 1649.44, "max": 2279.35, "n": 3 }, "median_tpot_ms": { "median": 177.46, "min": 130.85, "max": 227.55, "n": 3 }, "p99_tpot_ms": { "median": 206.84, "min": 180.71, "max": 256.61, "n": 3 }, "p99_itl_ms": { "median": 928.93, "min": 830.22, "max": 929.09, "n": 3 }, "median_itl_ms": { "median": 136.76, "min": 104.43, "max": 168.13, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.01, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.41, "min": 0.39, "max": 0.44, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 86.0, "thermal_share": 0.008 }, "ttft_quantiles_ms": [ 46.0, 61.6, 496.8, 515.9, 550.7, 571.3, 583.0, 595.7, 602.8, 624.4, 638.0, 666.9, 706.4, 715.0, 722.8, 735.1, 750.0, 771.4, 789.7, 796.9, 817.8, 824.4, 837.6, 842.8, 846.7, 863.3, 884.1, 903.1, 908.5, 920.4, 927.1, 952.1, 966.1, 975.1, 988.0, 995.9, 1021.3, 1030.3, 1037.9, 1071.0, 1116.3, 1161.8, 1201.9, 1238.3, 1257.6, 1301.9, 1332.2, 1547.1, 1590.6, 1750.0, 2364.5 ], "tpot_quantiles_ms": [ 74.9, 89.8, 104.6, 107.1, 109.9, 116.5, 117.4, 119.0, 122.3, 125.6, 130.0, 133.3, 133.8, 137.5, 141.1, 143.0, 146.2, 149.6, 152.3, 158.4, 160.1, 161.5, 163.6, 167.2, 168.6, 172.7, 173.6, 174.6, 177.3, 178.7, 181.5, 184.4, 188.7, 192.9, 197.3, 198.2, 199.4, 201.4, 203.2, 206.2, 218.1, 221.5, 228.7, 230.8, 234.0, 235.6, 237.3, 239.2, 242.4, 250.7, 256.6 ] }, { "x": 1.0, "output_throughput": { "median": 106.8, "min": 93.57, "max": 111.44, "n": 3 }, "median_ttft_ms": { "median": 268.04, "min": 223.15, "max": 274.04, "n": 3 }, "p99_ttft_ms": { "median": 434.8, "min": 318.14, "max": 435.6, "n": 3 }, "median_tpot_ms": { "median": 168.66, "min": 133.66, "max": 186.62, "n": 3 }, "p99_tpot_ms": { "median": 191.21, "min": 155.77, "max": 206.16, "n": 3 }, "p99_itl_ms": { "median": 239.77, "min": 182.25, "max": 262.84, "n": 3 }, "median_itl_ms": { "median": 167.09, "min": 134.28, "max": 192.65, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.0, "max": 0.08, "n": 3 }, "request_throughput": { "median": 0.66, "min": 0.63, "max": 0.67, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 45.7, 46.5, 74.3, 85.0, 94.2, 106.2, 117.7, 126.0, 144.3, 165.4, 170.9, 181.6, 193.1, 200.4, 202.1, 204.1, 212.3, 221.4, 228.1, 230.1, 237.9, 241.4, 243.7, 246.7, 251.9, 257.4, 260.7, 265.1, 271.0, 272.4, 273.6, 278.2, 286.6, 291.3, 294.5, 297.3, 298.6, 307.6, 313.7, 316.0, 322.1, 328.7, 339.5, 352.6, 361.6, 372.8, 391.4, 398.5, 403.2, 422.9, 447.9 ], "tpot_quantiles_ms": [ 66.2, 89.5, 98.0, 110.2, 114.8, 120.0, 126.1, 129.9, 130.8, 131.8, 133.9, 136.1, 137.6, 139.5, 140.6, 141.5, 141.9, 142.9, 143.4, 146.3, 146.7, 151.0, 152.3, 154.3, 156.1, 160.9, 164.7, 165.6, 167.5, 169.1, 169.5, 171.9, 173.9, 175.0, 179.8, 180.8, 181.1, 182.5, 182.8, 183.9, 185.4, 187.6, 190.2, 191.5, 193.8, 196.8, 197.6, 199.2, 201.5, 203.8, 206.9 ] }, { "x": 2.0, "output_throughput": { "median": 129.29, "min": 123.14, "max": 136.24, "n": 3 }, "median_ttft_ms": { "median": 269.25, "min": 256.3, "max": 275.71, "n": 3 }, "p99_ttft_ms": { "median": 452.01, "min": 396.36, "max": 478.7, "n": 3 }, "median_tpot_ms": { "median": 196.92, "min": 176.74, "max": 213.51, "n": 3 }, "p99_tpot_ms": { "median": 222.45, "min": 195.2, "max": 237.15, "n": 3 }, "p99_itl_ms": { "median": 312.16, "min": 233.18, "max": 323.89, "n": 3 }, "median_itl_ms": { "median": 201.72, "min": 173.35, "max": 223.57, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.81, "min": 0.8, "max": 0.83, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 45.8, 47.2, 74.5, 85.3, 94.4, 107.7, 122.1, 129.9, 131.2, 146.9, 170.4, 183.8, 186.4, 191.7, 195.6, 197.5, 207.0, 214.2, 220.2, 228.6, 239.9, 246.5, 251.5, 255.7, 258.6, 266.5, 274.0, 277.5, 281.6, 283.6, 286.5, 291.9, 296.2, 307.0, 314.1, 322.8, 328.7, 329.0, 337.5, 341.9, 350.7, 353.5, 360.5, 369.9, 374.1, 378.2, 387.0, 402.5, 433.9, 441.3, 511.9 ], "tpot_quantiles_ms": [ 112.4, 129.2, 137.1, 144.6, 146.4, 158.2, 160.9, 166.2, 167.6, 169.1, 172.9, 175.6, 176.6, 178.0, 178.4, 179.6, 180.6, 182.6, 183.4, 185.2, 187.2, 187.6, 188.2, 190.1, 191.9, 192.8, 193.9, 194.4, 195.3, 196.4, 197.8, 199.4, 203.5, 206.5, 208.1, 210.9, 211.4, 212.4, 212.9, 213.5, 214.7, 217.6, 219.1, 220.9, 222.2, 222.5, 226.9, 229.0, 231.4, 232.1, 239.9 ] } ] }, "mixed-chunk": { "kv_pool_tokens": 236751, "points": [ { "x": 0.5, "output_throughput": { "median": 72.08, "min": 58.36, "max": 72.29, "n": 3 }, "median_ttft_ms": { "median": 909.81, "min": 734.54, "max": 925.13, "n": 3 }, "p99_ttft_ms": { "median": 1743.87, "min": 1528.92, "max": 1821.79, "n": 3 }, "median_tpot_ms": { "median": 148.43, "min": 111.3, "max": 175.47, "n": 3 }, "p99_tpot_ms": { "median": 173.88, "min": 141.62, "max": 209.29, "n": 3 }, "p99_itl_ms": { "median": 780.5, "min": 739.43, "max": 802.99, "n": 3 }, "median_itl_ms": { "median": 122.52, "min": 93.77, "max": 135.49, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.01, "max": 0.11, "n": 3 }, "request_throughput": { "median": 0.43, "min": 0.39, "max": 0.45, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2405.0, "min": 2398.0, "max": 2411.0, "n": 3 }, "max_temp_c": 89.0, "thermal_share": 0.109 }, "ttft_quantiles_ms": [ 43.1, 56.1, 514.4, 533.4, 542.0, 574.0, 585.8, 603.1, 611.5, 617.1, 636.2, 659.5, 668.7, 672.2, 689.6, 702.6, 725.9, 733.8, 762.9, 778.7, 787.5, 804.6, 807.8, 827.4, 862.1, 872.7, 888.6, 907.1, 909.8, 919.1, 935.8, 946.1, 957.8, 965.5, 980.5, 987.0, 993.1, 1005.7, 1019.9, 1053.0, 1068.8, 1162.5, 1195.3, 1219.4, 1273.3, 1340.2, 1401.8, 1520.5, 1632.9, 1698.3, 1906.9 ], "tpot_quantiles_ms": [ 69.4, 79.0, 91.8, 94.5, 95.1, 98.3, 100.2, 104.5, 106.6, 107.6, 111.2, 114.6, 116.0, 117.2, 122.1, 125.0, 126.7, 127.1, 127.9, 128.5, 130.8, 131.7, 133.7, 136.1, 138.2, 140.5, 142.0, 144.3, 146.7, 147.2, 149.2, 151.5, 152.1, 157.5, 163.0, 164.5, 165.0, 165.5, 166.6, 170.1, 173.3, 174.4, 176.1, 181.5, 187.5, 192.9, 194.3, 195.8, 197.4, 200.1, 213.4 ] }, { "x": 1.0, "output_throughput": { "median": 112.61, "min": 96.98, "max": 117.3, "n": 3 }, "median_ttft_ms": { "median": 362.05, "min": 292.36, "max": 378.59, "n": 3 }, "p99_ttft_ms": { "median": 481.25, "min": 402.56, "max": 534.86, "n": 3 }, "median_tpot_ms": { "median": 145.68, "min": 117.77, "max": 158.18, "n": 3 }, "p99_tpot_ms": { "median": 159.9, "min": 134.97, "max": 170.05, "n": 3 }, "p99_itl_ms": { "median": 177.4, "min": 147.44, "max": 191.74, "n": 3 }, "median_itl_ms": { "median": 148.02, "min": 120.15, "max": 159.43, "n": 3 }, "goodput_rps": { "median": 0.03, "min": 0.0, "max": 0.1, "n": 3 }, "request_throughput": { "median": 0.69, "min": 0.65, "max": 0.71, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 81.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 44.1, 45.7, 125.8, 143.6, 178.5, 209.6, 233.1, 236.7, 243.6, 259.4, 262.7, 271.0, 277.1, 281.6, 286.1, 288.9, 295.0, 300.3, 306.5, 308.8, 314.2, 316.6, 321.3, 327.6, 337.8, 342.7, 346.5, 354.3, 356.7, 359.8, 363.4, 365.0, 373.7, 377.9, 378.7, 381.3, 388.7, 391.4, 397.6, 398.2, 401.9, 413.7, 428.4, 435.5, 445.7, 469.7, 476.8, 481.9, 487.8, 509.0, 544.5 ], "tpot_quantiles_ms": [ 55.2, 80.1, 87.3, 98.3, 103.7, 104.9, 109.3, 113.8, 114.0, 116.6, 117.4, 117.7, 118.5, 119.2, 120.4, 121.2, 123.5, 124.5, 125.0, 128.5, 129.6, 131.4, 133.4, 134.9, 137.0, 139.6, 142.8, 144.5, 145.4, 145.5, 145.7, 148.5, 150.6, 151.2, 152.3, 153.7, 154.5, 155.5, 156.0, 157.0, 157.5, 158.5, 160.2, 160.9, 163.2, 164.6, 165.8, 167.1, 167.9, 168.8, 170.4 ] }, { "x": 2.0, "output_throughput": { "median": 138.09, "min": 130.56, "max": 145.3, "n": 3 }, "median_ttft_ms": { "median": 415.63, "min": 368.64, "max": 415.72, "n": 3 }, "p99_ttft_ms": { "median": 617.65, "min": 495.94, "max": 628.19, "n": 3 }, "median_tpot_ms": { "median": 175.43, "min": 153.4, "max": 189.85, "n": 3 }, "p99_tpot_ms": { "median": 200.89, "min": 171.8, "max": 211.16, "n": 3 }, "p99_itl_ms": { "median": 220.97, "min": 183.22, "max": 226.26, "n": 3 }, "median_itl_ms": { "median": 181.73, "min": 154.16, "max": 192.89, "n": 3 }, "goodput_rps": { "median": 0.02, "min": 0.0, "max": 0.02, "n": 3 }, "request_throughput": { "median": 0.87, "min": 0.85, "max": 0.88, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 80.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 42.9, 46.6, 116.2, 154.8, 176.1, 219.3, 233.8, 243.1, 273.5, 284.6, 288.6, 294.2, 299.2, 314.9, 324.2, 326.5, 332.4, 340.4, 345.0, 354.0, 365.4, 372.9, 376.0, 388.3, 390.4, 392.4, 399.0, 405.1, 411.8, 413.8, 418.8, 424.1, 424.9, 433.1, 441.5, 446.7, 452.2, 456.8, 462.6, 476.4, 481.3, 484.9, 497.0, 504.9, 521.7, 528.0, 544.5, 550.6, 584.2, 607.5, 646.6 ], "tpot_quantiles_ms": [ 88.8, 109.8, 119.0, 122.7, 130.6, 138.8, 143.8, 147.2, 150.0, 150.9, 151.5, 152.5, 155.0, 156.2, 157.2, 157.8, 158.1, 159.3, 160.8, 162.1, 163.0, 164.6, 166.0, 168.3, 170.5, 171.5, 172.4, 173.8, 175.2, 175.5, 176.6, 177.8, 178.0, 179.7, 182.1, 184.3, 184.8, 187.5, 189.7, 190.1, 191.1, 193.5, 195.1, 195.9, 197.3, 199.9, 200.7, 201.7, 204.0, 205.4, 213.8 ] } ] }, "graphs-on-b": { "kv_pool_tokens": 251910, "points": [ { "x": 0.5, "output_throughput": { "median": 70.86, "min": 58.08, "max": 71.08, "n": 3 }, "median_ttft_ms": { "median": 841.12, "min": 807.48, "max": 914.14, "n": 3 }, "p99_ttft_ms": { "median": 1692.68, "min": 1675.46, "max": 2123.3, "n": 3 }, "median_tpot_ms": { "median": 170.07, "min": 125.45, "max": 205.85, "n": 3 }, "p99_tpot_ms": { "median": 196.34, "min": 164.52, "max": 234.8, "n": 3 }, "p99_itl_ms": { "median": 922.13, "min": 811.41, "max": 924.51, "n": 3 }, "median_itl_ms": { "median": 130.39, "min": 99.05, "max": 150.22, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.01, "max": 0.04, "n": 3 }, "request_throughput": { "median": 0.42, "min": 0.39, "max": 0.45, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 87.0, "thermal_share": 0.028 }, "ttft_quantiles_ms": [ 44.9, 57.9, 510.4, 543.5, 562.2, 585.2, 595.3, 604.3, 619.4, 632.7, 640.3, 662.2, 684.7, 703.3, 712.8, 730.4, 753.7, 756.1, 763.3, 784.5, 796.4, 811.4, 821.1, 825.3, 843.3, 845.5, 857.9, 869.9, 879.4, 902.9, 912.2, 920.8, 939.8, 946.9, 956.2, 965.1, 978.6, 983.2, 1013.0, 1019.3, 1048.0, 1111.6, 1157.4, 1196.6, 1247.5, 1293.8, 1400.0, 1529.7, 1596.6, 1715.9, 2264.6 ], "tpot_quantiles_ms": [ 72.3, 82.1, 97.9, 100.6, 104.0, 107.0, 112.5, 113.9, 114.3, 116.7, 124.7, 126.3, 128.5, 131.0, 135.3, 137.4, 138.8, 141.7, 144.3, 146.3, 147.0, 149.7, 150.8, 156.7, 161.4, 161.9, 166.3, 167.2, 168.7, 169.2, 170.2, 173.2, 175.2, 182.1, 183.2, 185.5, 186.3, 187.9, 189.1, 190.5, 195.2, 200.2, 209.0, 212.1, 214.7, 216.6, 221.9, 223.4, 224.6, 230.9, 234.8 ] }, { "x": 1.0, "output_throughput": { "median": 108.84, "min": 95.61, "max": 113.69, "n": 3 }, "median_ttft_ms": { "median": 243.63, "min": 222.37, "max": 287.29, "n": 3 }, "p99_ttft_ms": { "median": 409.45, "min": 355.32, "max": 445.83, "n": 3 }, "median_tpot_ms": { "median": 161.9, "min": 130.42, "max": 179.03, "n": 3 }, "p99_tpot_ms": { "median": 182.87, "min": 149.81, "max": 197.9, "n": 3 }, "p99_itl_ms": { "median": 219.39, "min": 178.76, "max": 235.68, "n": 3 }, "median_itl_ms": { "median": 163.19, "min": 129.39, "max": 183.48, "n": 3 }, "goodput_rps": { "median": 0.01, "min": 0.0, "max": 0.08, "n": 3 }, "request_throughput": { "median": 0.67, "min": 0.64, "max": 0.68, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 77.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 45.9, 48.0, 80.6, 96.1, 106.2, 107.6, 111.1, 130.3, 145.3, 152.7, 175.2, 178.6, 185.5, 195.5, 199.1, 203.0, 211.4, 214.6, 218.0, 222.6, 225.1, 227.8, 231.0, 233.7, 238.6, 243.6, 250.2, 254.1, 259.5, 263.8, 265.6, 268.7, 275.5, 288.7, 294.4, 297.6, 299.2, 303.2, 308.4, 315.1, 320.2, 321.7, 326.0, 337.2, 344.8, 347.5, 369.2, 379.4, 395.8, 420.1, 451.6 ], "tpot_quantiles_ms": [ 62.2, 86.1, 92.6, 106.0, 109.2, 114.7, 120.0, 125.3, 127.4, 128.5, 129.8, 130.8, 131.4, 132.8, 134.3, 134.7, 136.1, 136.9, 137.9, 138.3, 139.3, 145.1, 147.0, 149.2, 150.1, 153.0, 158.0, 159.4, 160.6, 161.9, 162.4, 165.4, 167.2, 167.8, 171.0, 173.0, 174.2, 175.0, 175.2, 176.8, 178.3, 179.5, 182.1, 183.0, 185.5, 189.2, 190.4, 193.3, 194.3, 196.3, 198.0 ] }, { "x": 2.0, "output_throughput": { "median": 133.22, "min": 124.75, "max": 139.78, "n": 3 }, "median_ttft_ms": { "median": 238.32, "min": 235.6, "max": 282.7, "n": 3 }, "p99_ttft_ms": { "median": 451.51, "min": 387.27, "max": 457.62, "n": 3 }, "median_tpot_ms": { "median": 189.75, "min": 170.87, "max": 206.34, "n": 3 }, "p99_tpot_ms": { "median": 217.96, "min": 190.47, "max": 225.32, "n": 3 }, "p99_itl_ms": { "median": 260.62, "min": 233.67, "max": 285.54, "n": 3 }, "median_itl_ms": { "median": 196.02, "min": 167.41, "max": 216.1, "n": 3 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 3 }, "request_throughput": { "median": 0.84, "min": 0.82, "max": 0.84, "n": 3 }, "completed": [ 48, 48, 48 ], "gpu": { "median_sm_mhz": { "median": 2411.0, "min": 2411.0, "max": 2411.0, "n": 3 }, "max_temp_c": 76.0, "thermal_share": 0.0 }, "ttft_quantiles_ms": [ 44.5, 48.9, 67.2, 83.5, 97.1, 109.9, 129.5, 132.1, 136.2, 156.6, 176.3, 180.4, 184.7, 189.3, 192.0, 196.3, 201.5, 204.3, 210.0, 215.2, 221.3, 225.8, 232.8, 235.8, 240.8, 243.6, 246.7, 250.8, 255.3, 258.3, 262.4, 269.4, 276.6, 281.1, 294.3, 304.8, 306.8, 309.8, 312.8, 324.6, 348.2, 355.7, 361.2, 367.4, 372.6, 382.7, 395.6, 405.3, 419.8, 445.5, 466.1 ], "tpot_quantiles_ms": [ 102.0, 122.4, 131.0, 136.7, 143.0, 149.5, 157.4, 159.8, 161.1, 164.6, 167.3, 168.0, 170.3, 172.1, 173.2, 173.3, 174.5, 176.6, 178.0, 178.4, 179.1, 180.1, 180.9, 184.0, 184.7, 185.7, 187.8, 188.5, 188.8, 190.1, 191.3, 193.1, 195.8, 199.5, 201.4, 203.0, 203.6, 204.9, 205.8, 206.5, 207.0, 207.9, 210.7, 211.8, 214.8, 215.6, 218.4, 220.2, 221.2, 222.5, 226.9 ] } ] } } }