{ "profile": "l4-gpu1", "slo": { "ttft_ms": 1000.0, "tpot_ms": 100.0 }, "environment": null, "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 ], "load": { "vllm-fast-b": { "kv_pool_tokens": 2928, "points": [ { "x": 1.0, "output_throughput": { "median": 3.98, "min": 3.94, "max": 4.01, "n": 2 }, "median_ttft_ms": { "median": 1437.17, "min": 1432.09, "max": 1442.26, "n": 2 }, "p99_ttft_ms": { "median": 1536.41, "min": 1456.27, "max": 1616.55, "n": 2 }, "median_tpot_ms": { "median": 241.7, "min": 240.36, "max": 243.03, "n": 2 }, "p99_tpot_ms": { "median": 250.83, "min": 244.78, "max": 256.88, "n": 2 }, "p99_itl_ms": { "median": 274.22, "min": 255.01, "max": 293.44, "n": 2 }, "median_itl_ms": { "median": 241.0, "min": 240.67, "max": 241.32, "n": 2 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 2 }, "request_throughput": { "median": 0.03, "min": 0.03, "max": 0.03, "n": 2 }, "completed": [ 16, 16 ], "gpu": { "median_sm_mhz": { "median": 375.0, "min": 375.0, "max": 375.0, "n": 2 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 277.9, 277.9, 280.4, 280.4, 1138.2, 1223.6, 1223.6, 1309.9, 1309.9, 1379.9, 1391.1, 1391.1, 1410.9, 1411.5, 1411.5, 1413.8, 1413.8, 1415.4, 1417.8, 1417.8, 1422.0, 1429.0, 1429.0, 1435.2, 1435.2, 1437.2, 1438.9, 1438.9, 1439.8, 1439.8, 1440.6, 1443.1, 1443.1, 1443.9, 1444.7, 1444.7, 1449.9, 1449.9, 1452.2, 1453.1, 1453.1, 1456.8, 1467.5, 1467.5, 1485.1, 1485.1, 1492.0, 1543.8, 1543.8, 1605.6, 1618.5 ], "tpot_quantiles_ms": [ 238.0, 238.0, 238.5, 238.5, 239.2, 239.3, 239.3, 239.6, 239.6, 239.8, 239.9, 239.9, 240.0, 240.1, 240.1, 240.2, 240.2, 240.3, 240.3, 240.3, 240.4, 240.7, 240.7, 240.8, 240.8, 240.8, 241.0, 241.0, 241.7, 241.7, 241.8, 242.3, 242.3, 242.3, 242.3, 242.3, 242.8, 242.8, 243.3, 243.4, 243.4, 245.2, 246.6, 246.6, 246.9, 246.9, 247.6, 253.2, 253.2, 253.6, 257.5 ] }, { "x": 4.0, "output_throughput": { "median": 13.02, "min": 12.97, "max": 13.06, "n": 2 }, "median_ttft_ms": { "median": 4944.93, "min": 4764.24, "max": 5125.62, "n": 2 }, "p99_ttft_ms": { "median": 5342.51, "min": 5265.54, "max": 5419.48, "n": 2 }, "median_tpot_ms": { "median": 272.6, "min": 270.11, "max": 275.08, "n": 2 }, "p99_tpot_ms": { "median": 303.82, "min": 302.11, "max": 305.54, "n": 2 }, "p99_itl_ms": { "median": 290.42, "min": 286.11, "max": 294.73, "n": 2 }, "median_itl_ms": { "median": 271.87, "min": 271.19, "max": 272.56, "n": 2 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 2 }, "request_throughput": { "median": 0.1, "min": 0.1, "max": 0.1, "n": 2 }, "completed": [ 16, 16 ], "gpu": { "median_sm_mhz": { "median": 427.5, "min": 420.0, "max": 435.0, "n": 2 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 276.5, 276.5, 287.4, 287.4, 1503.1, 1558.9, 1558.9, 1560.4, 1560.4, 1674.5, 1690.9, 1690.9, 1729.6, 3997.6, 3997.6, 3998.2, 3998.2, 3998.6, 4143.7, 4143.7, 4144.3, 4144.7, 4144.7, 4763.9, 4763.9, 4764.6, 4765.5, 4765.5, 5125.4, 5125.4, 5125.8, 5127.2, 5127.2, 5257.8, 5258.3, 5258.3, 5259.6, 5259.6, 5264.0, 5264.6, 5264.6, 5265.7, 5380.6, 5380.6, 5381.1, 5381.1, 5382.7, 5417.8, 5417.8, 5418.8, 5419.6 ], "tpot_quantiles_ms": [ 266.9, 266.9, 266.9, 266.9, 266.9, 267.6, 267.6, 267.6, 267.6, 267.6, 268.1, 268.1, 268.1, 268.1, 268.1, 268.2, 268.2, 268.2, 268.2, 268.2, 270.1, 270.1, 270.1, 270.1, 270.1, 274.8, 274.8, 274.8, 274.8, 274.8, 275.1, 275.1, 275.1, 275.1, 276.8, 276.8, 276.8, 276.8, 276.8, 293.3, 293.3, 295.1, 296.2, 296.2, 296.6, 296.6, 297.2, 303.0, 303.0, 305.4, 305.6 ] } ] }, "vllm-ngram": { "kv_pool_tokens": 2928, "points": [ { "x": 1.0, "output_throughput": { "median": 7.39, "min": 7.3, "max": 7.49, "n": 2 }, "median_ttft_ms": { "median": 1428.05, "min": 1420.77, "max": 1435.34, "n": 2 }, "p99_ttft_ms": { "median": 1659.74, "min": 1615.52, "max": 1703.96, "n": 2 }, "median_tpot_ms": { "median": 76.58, "min": 72.51, "max": 80.66, "n": 2 }, "p99_tpot_ms": { "median": 256.21, "min": 248.39, "max": 264.03, "n": 2 }, "p99_itl_ms": { "median": 284.06, "min": 276.87, "max": 291.26, "n": 2 }, "median_itl_ms": { "median": 265.39, "min": 265.32, "max": 265.47, "n": 2 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 2 }, "request_throughput": { "median": 0.06, "min": 0.06, "max": 0.06, "n": 2 }, "completed": [ 16, 16 ], "gpu": { "median_sm_mhz": { "median": 412.5, "min": 405.0, "max": 420.0, "n": 2 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 270.9, 270.9, 272.5, 272.5, 1200.3, 1202.3, 1202.3, 1202.5, 1202.5, 1212.6, 1217.7, 1217.7, 1373.3, 1395.4, 1395.4, 1403.5, 1403.5, 1414.1, 1419.3, 1419.3, 1419.7, 1427.4, 1427.4, 1428.6, 1428.6, 1431.4, 1435.1, 1435.1, 1435.6, 1435.6, 1435.8, 1436.2, 1436.2, 1436.7, 1436.8, 1436.8, 1445.6, 1445.6, 1451.6, 1504.1, 1504.1, 1504.5, 1511.6, 1511.6, 1517.6, 1517.6, 1526.4, 1551.5, 1551.5, 1626.8, 1748.5 ], "tpot_quantiles_ms": [ 257.6, 257.6, 257.7, 257.7, 259.2, 260.4, 260.4, 261.0, 261.0, 262.2, 262.8, 262.8, 264.9, 265.7, 265.7, 266.9, 266.9, 267.0, 267.3, 267.3, 267.8, 268.2, 268.2, 268.5, 268.5, 268.6, 269.5, 269.5, 270.5, 270.5, 270.7, 270.8, 270.8, 270.9, 271.0, 271.0, 271.0, 271.0, 271.1, 271.4, 271.4, 271.8, 271.8, 271.8, 272.0, 272.0, 272.8, 276.3, 276.3, 277.4, 278.5 ] }, { "x": 4.0, "output_throughput": { "median": 19.22, "min": 18.75, "max": 19.69, "n": 2 }, "median_ttft_ms": { "median": 1924.53, "min": 1858.69, "max": 1990.37, "n": 2 }, "p99_ttft_ms": { "median": 3859.52, "min": 3818.24, "max": 3900.8, "n": 2 }, "median_tpot_ms": { "median": 101.25, "min": 95.01, "max": 107.49, "n": 2 }, "p99_tpot_ms": { "median": 349.44, "min": 347.63, "max": 351.25, "n": 2 }, "p99_itl_ms": { "median": 1691.13, "min": 1668.81, "max": 1713.46, "n": 2 }, "median_itl_ms": { "median": 283.7, "min": 282.9, "max": 284.51, "n": 2 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 2 }, "request_throughput": { "median": 0.15, "min": 0.15, "max": 0.15, "n": 2 }, "completed": [ 16, 16 ], "gpu": { "median_sm_mhz": { "median": 442.5, "min": 435.0, "max": 450.0, "n": 2 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 264.1, 264.1, 275.3, 275.3, 1673.4, 1706.4, 1706.4, 1733.2, 1733.2, 1746.4, 1746.7, 1746.7, 1749.2, 1750.1, 1750.1, 1754.2, 1754.2, 1762.5, 1864.2, 1864.2, 1894.3, 1932.5, 1932.5, 1954.9, 1954.9, 1973.6, 1982.8, 1982.8, 1998.0, 1998.0, 1998.8, 2048.3, 2048.3, 2149.8, 2158.1, 2158.1, 2777.3, 2777.3, 2858.5, 2970.3, 2970.3, 3729.2, 3817.4, 3817.4, 3817.9, 3817.9, 3818.3, 3900.2, 3900.2, 3900.8, 3900.8 ], "tpot_quantiles_ms": [ 263.8, 263.8, 269.8, 269.8, 278.0, 281.9, 281.9, 294.1, 294.1, 295.7, 301.5, 301.5, 308.8, 316.2, 316.2, 319.3, 319.3, 320.7, 326.9, 326.9, 333.2, 333.5, 333.5, 334.0, 334.0, 335.8, 341.1, 341.1, 341.5, 341.5, 342.7, 345.5, 345.5, 348.6, 352.3, 352.3, 357.2, 357.2, 357.4, 358.1, 358.1, 361.2, 368.2, 368.2, 368.2, 368.2, 369.8, 380.6, 380.6, 386.7, 393.7 ] } ] }, "vllm-no-async": { "kv_pool_tokens": 2928, "points": [ { "x": 1.0, "output_throughput": { "median": 3.87, "min": 3.8, "max": 3.93, "n": 2 }, "median_ttft_ms": { "median": 1466.72, "min": 1430.97, "max": 1502.47, "n": 2 }, "p99_ttft_ms": { "median": 1783.18, "min": 1699.69, "max": 1866.67, "n": 2 }, "median_tpot_ms": { "median": 249.58, "min": 245.33, "max": 253.83, "n": 2 }, "p99_tpot_ms": { "median": 259.08, "min": 249.41, "max": 268.74, "n": 2 }, "p99_itl_ms": { "median": 282.54, "min": 257.85, "max": 307.24, "n": 2 }, "median_itl_ms": { "median": 249.77, "min": 245.75, "max": 253.79, "n": 2 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 2 }, "request_throughput": { "median": 0.03, "min": 0.03, "max": 0.03, "n": 2 }, "completed": [ 16, 16 ], "gpu": { "median_sm_mhz": { "median": 367.5, "min": 360.0, "max": 375.0, "n": 2 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 279.9, 279.9, 287.2, 287.2, 1192.8, 1260.3, 1260.3, 1301.6, 1301.6, 1321.3, 1406.9, 1406.9, 1418.0, 1419.7, 1419.7, 1427.2, 1427.2, 1428.2, 1433.8, 1433.8, 1435.6, 1438.9, 1438.9, 1448.9, 1448.9, 1451.3, 1451.4, 1451.4, 1458.3, 1458.3, 1470.6, 1478.3, 1478.3, 1483.0, 1491.8, 1491.8, 1499.4, 1499.4, 1505.5, 1512.5, 1512.5, 1517.3, 1529.1, 1529.1, 1652.6, 1652.6, 1742.3, 1751.4, 1751.4, 1787.1, 1880.7 ], "tpot_quantiles_ms": [ 239.3, 239.3, 240.6, 240.6, 240.8, 243.5, 243.5, 243.7, 243.7, 243.7, 243.9, 243.9, 243.9, 244.2, 244.2, 244.4, 244.4, 245.2, 245.4, 245.4, 245.5, 246.5, 246.5, 247.3, 247.3, 248.0, 248.6, 248.6, 249.3, 249.3, 249.3, 249.4, 249.4, 251.6, 251.9, 251.9, 252.2, 252.2, 252.8, 254.8, 254.8, 255.3, 255.7, 255.7, 255.8, 255.8, 259.4, 262.6, 262.6, 264.3, 269.5 ] }, { "x": 4.0, "output_throughput": { "median": 12.85, "min": 12.85, "max": 12.85, "n": 1 }, "median_ttft_ms": { "median": 4690.01, "min": 4690.01, "max": 4690.01, "n": 1 }, "p99_ttft_ms": { "median": 5302.54, "min": 5302.54, "max": 5302.54, "n": 1 }, "median_tpot_ms": { "median": 277.22, "min": 277.22, "max": 277.22, "n": 1 }, "p99_tpot_ms": { "median": 306.43, "min": 306.43, "max": 306.43, "n": 1 }, "p99_itl_ms": { "median": 309.49, "min": 309.49, "max": 309.49, "n": 1 }, "median_itl_ms": { "median": 276.84, "min": 276.84, "max": 276.84, "n": 1 }, "goodput_rps": { "median": 0.0, "min": 0.0, "max": 0.0, "n": 1 }, "request_throughput": { "median": 0.1, "min": 0.1, "max": 0.1, "n": 1 }, "completed": [ 16 ], "gpu": { "median_sm_mhz": { "median": 450.0, "min": 450.0, "max": 450.0, "n": 1 }, "max_temp_c": 88.0, "thermal_share": 1.0 }, "ttft_quantiles_ms": [ 288.0, 288.0, 288.0, 288.0, 1435.6, 1435.6, 1435.6, 1495.2, 1495.2, 1495.2, 1577.5, 1577.5, 1577.5, 1577.5, 3903.2, 3903.2, 3903.2, 3903.8, 3903.8, 3903.8, 3904.2, 3904.2, 3904.2, 3904.2, 4689.7, 4689.7, 4689.7, 4690.4, 4690.4, 4690.4, 4691.4, 4691.4, 4691.4, 4691.4, 5206.1, 5206.1, 5206.1, 5206.5, 5206.5, 5206.5, 5207.8, 5207.8, 5207.8, 5207.8, 5301.0, 5301.0, 5301.0, 5301.4, 5301.4, 5301.4, 5302.7 ], "tpot_quantiles_ms": [ 271.9, 271.9, 271.9, 271.9, 271.9, 271.9, 271.9, 271.9, 271.9, 271.9, 275.8, 275.8, 275.8, 275.8, 275.8, 275.8, 275.8, 275.8, 275.8, 275.8, 277.2, 277.2, 277.2, 277.2, 277.2, 277.2, 277.2, 277.2, 277.2, 277.2, 279.8, 279.8, 279.8, 279.8, 279.8, 279.8, 279.8, 279.8, 279.8, 279.8, 298.3, 298.3, 298.3, 298.3, 305.0, 305.0, 305.0, 305.4, 305.4, 305.4, 306.6 ] } ] } } }