## single stream (ms) A-live b1_3 n= 40 e2e p50 186.9 [178.2,197.3] p90 221.1 p99 230.2 max 232.6 | server p50 nan p90 nan p99 nan | http overhead p50 nan A-live b3_8 n= 40 e2e p50 307.5 [274.5,329.2] p90 412.5 p99 442.1 max 454.0 | server p50 nan p90 nan p99 nan | http overhead p50 nan A-live b8_20 n= 40 e2e p50 625.7 [546.5,678.5] p90 861.8 p99 950.5 max 955.6 | server p50 nan p90 nan p99 nan | http overhead p50 nan ab-a1 b1_3 n=120 e2e p50 144.2 [141.2,151.5] p90 170.4 p99 183.6 max 190.5 | server p50 142.2 p90 168.5 p99 181.7 | http overhead p50 1.9 ab-a1 b3_8 n=120 e2e p50 259.8 [250.6,279.0] p90 370.1 p99 383.9 max 396.0 | server p50 257.8 p90 367.9 p99 381.8 | http overhead p50 2.0 ab-a1 b8_20 n=120 e2e p50 564.7 [537.4,599.1] p90 814.4 p99 896.0 max 926.6 | server p50 562.5 p90 812.0 p99 893.5 | http overhead p50 2.2 ab-a1 b20_60 n=120 e2e p50 2026.8 [1773.8,2130.3] p90 2634.4 p99 2768.3 max 2918.0 | server p50 2021.9 p90 2628.0 p99 2763.1 | http overhead p50 5.0 ab-a2 b1_3 n=120 e2e p50 149.6 [144.8,155.3] p90 176.1 p99 206.5 max 231.3 | server p50 147.7 p90 174.1 p99 204.2 | http overhead p50 1.8 ab-a2 b3_8 n=120 e2e p50 265.1 [253.5,283.9] p90 374.2 p99 394.3 max 428.2 | server p50 263.0 p90 372.1 p99 392.3 | http overhead p50 1.9 ab-a2 b8_20 n=120 e2e p50 567.6 [526.2,604.2] p90 814.2 p99 901.5 max 915.0 | server p50 565.6 p90 811.9 p99 898.9 | http overhead p50 2.1 ab-a2 b20_60 n=120 e2e p50 2010.0 [1788.5,2146.1] p90 2717.1 p99 2826.1 max 2888.5 | server p50 2004.8 p90 2709.4 p99 2818.2 | http overhead p50 4.9 ab-a50 b1_3 n=120 e2e p50 197.0 [193.7,204.2] p90 224.0 p99 233.4 max 250.7 | server p50 194.6 p90 222.2 p99 231.1 | http overhead p50 1.8 ab-a50 b3_8 n=120 e2e p50 313.3 [300.9,332.2] p90 424.4 p99 454.8 max 462.8 | server p50 311.3 p90 422.4 p99 452.5 | http overhead p50 2.0 ab-a50 b8_20 n=120 e2e p50 615.3 [576.1,648.8] p90 861.4 p99 943.2 max 948.8 | server p50 613.0 p90 859.1 p99 940.8 | http overhead p50 2.2 ab-a50 b20_60 n=120 e2e p50 2072.3 [1835.1,2252.7] p90 2678.1 p99 2828.9 max 2864.5 | server p50 2067.0 p90 2671.9 p99 2822.7 | http overhead p50 4.9 ab-af32 b1_3 n=120 e2e p50 38.1 [36.8,39.1] p90 41.0 p99 43.5 max 47.6 | server p50 36.5 p90 39.4 p99 41.8 | http overhead p50 1.5 ab-af32 b3_8 n=120 e2e p50 46.7 [45.9,47.7] p90 53.1 p99 57.1 max 77.3 | server p50 45.0 p90 51.3 p99 55.3 | http overhead p50 1.7 ab-af32 b8_20 n=120 e2e p50 67.0 [65.3,71.3] p90 80.4 p99 87.2 max 95.3 | server p50 65.0 p90 78.2 p99 84.5 | http overhead p50 1.9 ab-af32 b20_60 n=120 e2e p50 167.0 [157.2,175.1] p90 219.2 p99 231.1 max 233.1 | server p50 162.1 p90 214.1 p99 225.9 | http overhead p50 4.6 ab-apl b1_3 n=120 e2e p50 149.9 [146.9,157.7] p90 175.8 p99 189.6 max 191.6 | server p50 nan p90 nan p99 nan | http overhead p50 nan ab-apl b3_8 n=120 e2e p50 265.2 [253.8,287.0] p90 378.3 p99 406.3 max 425.4 | server p50 nan p90 nan p99 nan | http overhead p50 nan ab-apl b8_20 n=120 e2e p50 577.5 [547.2,610.4] p90 815.7 p99 927.9 max 946.4 | server p50 nan p90 nan p99 nan | http overhead p50 nan ab-apl b20_60 n=120 e2e p50 2043.4 [1819.7,2134.5] p90 2742.3 p99 2837.1 max 2917.1 | server p50 nan p90 nan p99 nan | http overhead p50 nan ab-at16 b1_3 n=120 e2e p50 85.0 [82.7,88.6] p90 97.6 p99 104.4 max 105.7 | server p50 83.1 p90 95.7 p99 100.4 | http overhead p50 1.9 ab-at16 b3_8 n=120 e2e p50 142.6 [135.6,151.9] p90 185.3 p99 196.8 max 198.4 | server p50 140.6 p90 183.0 p99 194.6 | http overhead p50 2.0 ab-at16 b8_20 n=120 e2e p50 318.0 [298.3,332.4] p90 437.7 p99 506.7 max 516.9 | server p50 315.7 p90 435.2 p99 504.3 | http overhead p50 2.2 ab-at16 b20_60 n=120 e2e p50 1171.3 [1056.1,1275.6] p90 1545.7 p99 1610.0 max 1630.2 | server p50 1167.1 p90 1540.7 p99 1603.2 | http overhead p50 4.7 ab-b16 b1_3 n=120 e2e p50 28.5 [28.2,28.6] p90 31.4 p99 55.4 max 63.6 | server p50 27.2 p90 30.2 p99 52.5 | http overhead p50 1.3 ab-b16 b3_8 n=120 e2e p50 32.4 [31.9,33.0] p90 35.9 p99 38.5 max 63.2 | server p50 31.0 p90 34.4 p99 37.1 | http overhead p50 1.4 ab-b16 b8_20 n=120 e2e p50 38.8 [38.1,39.3] p90 43.0 p99 45.7 max 75.7 | server p50 37.1 p90 41.1 p99 43.9 | http overhead p50 1.7 ab-b16 b20_60 n=120 e2e p50 66.6 [64.4,72.5] p90 82.5 p99 94.9 max 428.4 | server p50 62.4 p90 77.9 p99 85.9 | http overhead p50 4.0 ab-b32 b1_3 n=120 e2e p50 22.9 [22.7,23.2] p90 24.9 p99 29.4 max 60.7 | server p50 21.5 p90 23.4 p99 26.7 | http overhead p50 1.3 ab-b32 b3_8 n=120 e2e p50 26.6 [25.9,26.9] p90 29.2 p99 30.4 max 51.3 | server p50 25.2 p90 27.6 p99 29.0 | http overhead p50 1.4 ab-b32 b8_20 n=120 e2e p50 34.2 [33.3,35.1] p90 40.1 p99 61.9 max 238.1 | server p50 32.4 p90 37.9 p99 42.6 | http overhead p50 1.7 ab-b32 b20_60 n=120 e2e p50 76.7 [72.3,84.5] p90 97.9 p99 257.7 max 418.2 | server p50 72.2 p90 92.7 p99 96.4 | http overhead p50 4.3 ab-c b1_3 n=120 e2e p50 155.7 [150.8,161.8] p90 180.3 p99 197.2 max 220.9 | server p50 153.7 p90 178.2 p99 195.2 | http overhead p50 1.9 ab-c b3_8 n=120 e2e p50 271.1 [260.9,296.2] p90 389.1 p99 413.8 max 414.1 | server p50 269.1 p90 387.1 p99 411.6 | http overhead p50 2.0 ab-c b8_20 n=120 e2e p50 591.3 [557.3,626.8] p90 844.6 p99 941.6 max 951.6 | server p50 589.1 p90 840.8 p99 939.1 | http overhead p50 2.3 ab-c b20_60 n=120 e2e p50 2101.0 [1846.4,2203.2] p90 2795.4 p99 2942.6 max 2968.9 | server p50 2096.2 p90 2787.6 p99 2934.8 | http overhead p50 5.8 ab-cf16 b1_3 n=120 e2e p50 88.6 [88.1,89.1] p90 91.3 p99 115.8 max 131.5 | server p50 86.6 p90 89.4 p99 112.2 | http overhead p50 1.9 ab-cf16 b3_8 n=120 e2e p50 98.9 [97.6,99.7] p90 107.1 p99 111.2 max 126.5 | server p50 96.9 p90 105.0 p99 108.0 | http overhead p50 2.0 ab-cf16 b8_20 n=120 e2e p50 123.6 [122.5,127.9] p90 141.2 p99 148.1 max 150.8 | server p50 121.4 p90 138.7 p99 145.7 | http overhead p50 2.3 ab-cf16 b20_60 n=120 e2e p50 244.1 [230.9,253.7] p90 306.5 p99 317.7 max 323.2 | server p50 238.5 p90 300.1 p99 310.3 | http overhead p50 5.4 ab-cf32 b1_3 n=120 e2e p50 40.7 [39.6,41.5] p90 43.3 p99 66.5 max 73.8 | server p50 39.2 p90 41.5 p99 63.2 | http overhead p50 1.6 ab-cf32 b3_8 n=120 e2e p50 50.4 [49.6,51.8] p90 58.1 p99 62.9 max 63.1 | server p50 48.6 p90 56.2 p99 60.8 | http overhead p50 1.8 ab-cf32 b8_20 n=120 e2e p50 75.5 [73.7,79.0] p90 91.0 p99 97.8 max 99.7 | server p50 73.3 p90 88.6 p99 95.4 | http overhead p50 2.1 ab-cf32 b20_60 n=120 e2e p50 186.2 [173.7,203.1] p90 247.8 p99 260.2 max 261.4 | server p50 180.6 p90 241.5 p99 253.6 | http overhead p50 4.7 ab-d b1_3 n=120 e2e p50 149.5 [145.5,156.1] p90 173.9 p99 187.7 max 192.5 | server p50 147.5 p90 171.9 p99 185.7 | http overhead p50 2.0 ab-d b3_8 n=120 e2e p50 263.8 [249.8,285.1] p90 378.6 p99 390.0 max 395.3 | server p50 261.6 p90 376.2 p99 387.5 | http overhead p50 2.1 ab-d b8_20 n=120 e2e p50 564.9 [533.2,608.5] p90 805.5 p99 900.6 max 920.3 | server p50 562.6 p90 803.1 p99 897.6 | http overhead p50 2.3 ab-d b20_60 n=120 e2e p50 2003.5 [1769.0,2152.5] p90 2718.3 p99 2767.2 max 2793.0 | server p50 1998.0 p90 2711.9 p99 2760.8 | http overhead p50 5.1 ## paired median difference vs ab-a1 (e2e ms; 95% bootstrap CI) A-live b1_3 pairs= 40 diff +36.6 [ +34.3, +37.7] A-live b3_8 pairs= 40 diff +39.6 [ +38.2, +40.7] A-live b8_20 pairs= 40 diff +42.2 [ +36.2, +47.2] ab-a2 b1_3 pairs=120 diff +1.0 [ +0.7, +2.1] ab-a2 b3_8 pairs=120 diff +1.5 [ +0.9, +2.7] ab-a2 b8_20 pairs=120 diff +1.8 [ +0.4, +3.2] ab-a2 b20_60 pairs=120 diff +2.2 [ -3.8, +7.5] ab-a50 b1_3 pairs=120 diff +52.6 [ +52.1, +53.1] ab-a50 b3_8 pairs=120 diff +53.2 [ +52.5, +53.8] ab-a50 b8_20 pairs=120 diff +52.3 [ +50.8, +53.3] ab-a50 b20_60 pairs=120 diff +53.9 [ +50.6, +56.7] ab-af32 b1_3 pairs=120 diff -106.8 [ -114.2, -103.9] ab-af32 b3_8 pairs=120 diff -212.8 [ -234.2, -204.0] ab-af32 b8_20 pairs=120 diff -495.0 [ -528.9, -465.3] ab-af32 b20_60 pairs=120 diff -1858.7 [ -1962.3, -1616.2] ab-apl b1_3 pairs=120 diff +4.4 [ +3.9, +5.1] ab-apl b3_8 pairs=120 diff +5.8 [ +5.0, +6.8] ab-apl b8_20 pairs=120 diff +6.6 [ +4.9, +10.4] ab-apl b20_60 pairs=120 diff +16.8 [ +13.2, +22.3] ab-at16 b1_3 pairs=120 diff -60.8 [ -62.9, -58.3] ab-at16 b3_8 pairs=120 diff -119.0 [ -128.8, -112.7] ab-at16 b8_20 pairs=120 diff -248.7 [ -266.3, -231.7] ab-at16 b20_60 pairs=120 diff -844.0 [ -909.8, -749.0] ab-b16 b1_3 pairs=120 diff -114.8 [ -121.8, -111.6] ab-b16 b3_8 pairs=120 diff -229.0 [ -246.1, -218.0] ab-b16 b8_20 pairs=120 diff -526.7 [ -557.5, -490.8] ab-b16 b20_60 pairs=120 diff -1959.2 [ -2071.3, -1710.1] ab-b32 b1_3 pairs=120 diff -121.4 [ -128.4, -117.3] ab-b32 b3_8 pairs=120 diff -234.3 [ -252.9, -225.3] ab-b32 b8_20 pairs=120 diff -530.1 [ -564.6, -498.9] ab-b32 b20_60 pairs=120 diff -1947.1 [ -2059.2, -1686.0] ab-c b1_3 pairs=120 diff +7.7 [ +6.9, +8.3] ab-c b3_8 pairs=120 diff +12.2 [ +10.8, +14.2] ab-c b8_20 pairs=120 diff +22.0 [ +19.2, +24.3] ab-c b20_60 pairs=120 diff +70.9 [ +61.5, +80.7] ab-cf16 b1_3 pairs=120 diff -55.3 [ -61.2, -51.3] ab-cf16 b3_8 pairs=120 diff -160.3 [ -181.0, -152.1] ab-cf16 b8_20 pairs=120 diff -438.7 [ -472.0, -408.2] ab-cf16 b20_60 pairs=120 diff -1779.6 [ -1878.3, -1538.7] ab-cf32 b1_3 pairs=120 diff -103.8 [ -109.3, -101.4] ab-cf32 b3_8 pairs=120 diff -208.6 [ -228.4, -200.8] ab-cf32 b8_20 pairs=120 diff -488.1 [ -520.9, -460.5] ab-cf32 b20_60 pairs=120 diff -1838.5 [ -1935.3, -1594.5] ab-d b1_3 pairs=120 diff +2.9 [ +2.4, +3.4] ab-d b3_8 pairs=120 diff +3.5 [ +2.9, +4.2] ab-d b8_20 pairs=120 diff +2.9 [ +2.0, +4.7] ab-d b20_60 pairs=120 diff +4.3 [ +0.3, +10.1] ## concurrency 4 (ms) ab-a1 b1_3 n=40 err=0 e2e p50 568.6 p90 617.4 p99 646.6 | server p50 144.5 | 6.86 req/s, 15.3x realtime ab-a1 b20_60 n=40 err=0 e2e p50 8313.4 p90 12135.7 p99 13743.7 | server p50 2057.7 | 0.49 req/s, 21.2x realtime ab-a1 b3_8 n=40 err=0 e2e p50 1062.2 p90 1406.1 p99 1586.2 | server p50 269.7 | 3.68 req/s, 18.3x realtime ab-a1 b8_20 n=40 err=0 e2e p50 2303.8 p90 2745.2 p99 3574.6 | server p50 569.6 | 1.7 req/s, 20.3x realtime ab-a2 b1_3 n=40 err=0 e2e p50 561.1 p90 613.9 p99 638.7 | server p50 141.6 | 6.92 req/s, 15.5x realtime ab-a2 b20_60 n=40 err=0 e2e p50 7423.6 p90 12493.0 p99 15152.0 | server p50 2050.5 | 0.5 req/s, 21.3x realtime ab-a2 b3_8 n=40 err=0 e2e p50 1072.9 p90 1260.2 p99 1561.3 | server p50 267.7 | 3.73 req/s, 18.6x realtime ab-a2 b8_20 n=40 err=0 e2e p50 2294.5 p90 2734.5 p99 3556.0 | server p50 566.8 | 1.7 req/s, 20.4x realtime ab-a50 b1_3 n=40 err=0 e2e p50 773.7 p90 825.7 p99 857.5 | server p50 194.2 | 5.07 req/s, 11.3x realtime ab-a50 b20_60 n=40 err=0 e2e p50 8686.4 p90 12353.0 p99 16021.3 | server p50 2135.7 | 0.48 req/s, 20.5x realtime ab-a50 b3_8 n=40 err=0 e2e p50 1267.4 p90 1669.3 p99 1867.1 | server p50 320.0 | 3.1 req/s, 15.4x realtime ab-a50 b8_20 n=40 err=0 e2e p50 2531.8 p90 3318.8 p99 3462.2 | server p50 620.4 | 1.55 req/s, 18.6x realtime ab-af32 b1_3 n=40 err=0 e2e p50 137.6 p90 152.7 p99 158.4 | server p50 36.2 | 28.29 req/s, 63.3x realtime ab-af32 b20_60 n=40 err=0 e2e p50 661.7 p90 1119.5 p99 1284.2 | server p50 160.8 | 5.91 req/s, 253.7x realtime ab-af32 b3_8 n=40 err=0 e2e p50 181.3 p90 220.0 p99 263.7 | server p50 44.2 | 21.89 req/s, 109.0x realtime ab-af32 b8_20 n=40 err=0 e2e p50 264.5 p90 289.0 p99 348.8 | server p50 64.0 | 14.95 req/s, 179.2x realtime ab-apl b1_3 n=40 err=0 e2e p50 603.4 p90 658.4 p99 702.4 | server p50 nan | 6.47 req/s, 14.5x realtime ab-apl b20_60 n=40 err=0 e2e p50 8303.4 p90 13690.5 p99 15568.7 | server p50 nan | 0.49 req/s, 20.8x realtime ab-apl b3_8 n=40 err=0 e2e p50 1090.3 p90 1484.1 p99 1689.5 | server p50 nan | 3.59 req/s, 17.9x realtime ab-apl b8_20 n=40 err=0 e2e p50 2348.5 p90 3265.2 p99 3707.9 | server p50 nan | 1.64 req/s, 19.6x realtime ab-at16 b1_3 n=40 err=0 e2e p50 335.1 p90 359.1 p99 378.5 | server p50 83.7 | 11.72 req/s, 26.2x realtime ab-at16 b20_60 n=40 err=0 e2e p50 4646.5 p90 7005.3 p99 7903.2 | server p50 1165.4 | 0.86 req/s, 36.9x realtime ab-at16 b3_8 n=40 err=0 e2e p50 571.7 p90 653.7 p99 830.9 | server p50 144.8 | 6.99 req/s, 34.8x realtime ab-at16 b8_20 n=40 err=0 e2e p50 1299.3 p90 1530.2 p99 1788.6 | server p50 326.2 | 3.06 req/s, 36.6x realtime ab-b16 b1_3 n=40 err=0 e2e p50 105.9 p90 107.6 p99 132.5 | server p50 25.9 | 36.36 req/s, 81.3x realtime ab-b16 b20_60 n=40 err=0 e2e p50 253.3 p90 383.3 p99 437.2 | server p50 61.0 | 15.28 req/s, 655.9x realtime ab-b16 b3_8 n=40 err=0 e2e p50 122.4 p90 144.2 p99 173.9 | server p50 29.8 | 32.73 req/s, 162.9x realtime ab-b16 b8_20 n=40 err=0 e2e p50 146.5 p90 152.3 p99 189.2 | server p50 35.6 | 27.17 req/s, 325.8x realtime ab-b32 b1_3 n=40 err=0 e2e p50 84.6 p90 86.4 p99 113.4 | server p50 20.7 | 45.1 req/s, 100.9x realtime ab-b32 b20_60 n=40 err=0 e2e p50 308.1 p90 455.7 p99 538.7 | server p50 72.0 | 13.06 req/s, 560.6x realtime ab-b32 b3_8 n=40 err=0 e2e p50 98.4 p90 119.9 p99 140.3 | server p50 24.3 | 40.28 req/s, 200.5x realtime ab-b32 b8_20 n=40 err=0 e2e p50 128.8 p90 138.8 p99 167.2 | server p50 31.5 | 30.37 req/s, 364.1x realtime ab-c b1_3 n=40 err=0 e2e p50 636.2 p90 719.0 p99 720.2 | server p50 160.4 | 6.18 req/s, 13.8x realtime ab-c b20_60 n=40 err=0 e2e p50 8762.8 p90 12716.7 p99 14721.2 | server p50 2141.1 | 0.47 req/s, 20.4x realtime ab-c b3_8 n=40 err=0 e2e p50 1136.8 p90 1510.3 p99 1753.8 | server p50 284.6 | 3.45 req/s, 17.2x realtime ab-c b8_20 n=40 err=0 e2e p50 2414.6 p90 2882.7 p99 3714.5 | server p50 602.1 | 1.62 req/s, 19.4x realtime ab-cf16 b1_3 n=40 err=0 e2e p50 348.0 p90 355.3 p99 379.0 | server p50 86.5 | 11.53 req/s, 25.8x realtime ab-cf16 b20_60 n=40 err=0 e2e p50 1021.2 p90 1560.8 p99 1847.9 | server p50 239.8 | 4.0 req/s, 171.8x realtime ab-cf16 b3_8 n=40 err=0 e2e p50 390.7 p90 499.1 p99 558.5 | server p50 98.1 | 10.15 req/s, 50.5x realtime ab-cf16 b8_20 n=40 err=0 e2e p50 490.2 p90 522.2 p99 645.1 | server p50 121.1 | 8.06 req/s, 96.6x realtime ab-cf32 b1_3 n=40 err=0 e2e p50 148.4 p90 163.0 p99 165.8 | server p50 38.9 | 26.44 req/s, 59.1x realtime ab-cf32 b20_60 n=40 err=0 e2e p50 785.5 p90 1399.5 p99 1551.3 | server p50 186.5 | 5.08 req/s, 218.2x realtime ab-cf32 b3_8 n=40 err=0 e2e p50 199.4 p90 226.1 p99 285.8 | server p50 50.2 | 19.69 req/s, 98.0x realtime ab-cf32 b8_20 n=40 err=0 e2e p50 294.6 p90 328.6 p99 394.3 | server p50 72.1 | 13.3 req/s, 159.4x realtime ab-d b1_3 n=40 err=0 e2e p50 610.3 p90 658.2 p99 688.3 | server p50 151.4 | 6.49 req/s, 14.5x realtime ab-d b20_60 n=40 err=0 e2e p50 8643.5 p90 12435.2 p99 14957.1 | server p50 2062.9 | 0.49 req/s, 21.0x realtime ab-d b3_8 n=40 err=0 e2e p50 1104.4 p90 1462.3 p99 1633.0 | server p50 274.8 | 3.54 req/s, 17.6x realtime ab-d b8_20 n=40 err=0 e2e p50 2350.3 p90 2855.8 p99 3505.1 | server p50 577.1 | 1.65 req/s, 19.8x realtime ## first call per bin after cold start (ms) vs that arm's warm e2e p50 ab-a1 b1_3 first 223.4 warm p50 144.2 ab-a1 b3_8 first 357.1 warm p50 259.8 ab-a1 b8_20 first 648.5 warm p50 564.7 ab-a1 b20_60 first 2903.5 warm p50 2026.8 ab-a2 b1_3 first 216.3 warm p50 149.6 ab-a2 b3_8 first 293.6 warm p50 265.1 ab-a2 b8_20 first 693.1 warm p50 567.6 ab-a2 b20_60 first 2897.1 warm p50 2010.0 ab-a50 b1_3 first 261.0 warm p50 197.0 ab-a50 b3_8 first 358.0 warm p50 313.3 ab-a50 b8_20 first 727.0 warm p50 615.3 ab-a50 b20_60 first 2915.8 warm p50 2072.3 ab-apl b1_3 first 205.3 warm p50 149.9 ab-apl b3_8 first 291.3 warm p50 265.2 ab-apl b8_20 first 670.6 warm p50 577.5 ab-apl b20_60 first 2922.8 warm p50 2043.4 ab-at16 b1_3 first 129.0 warm p50 85.0 ab-at16 b3_8 first 159.7 warm p50 142.6 ab-at16 b8_20 first 324.9 warm p50 318.0 ab-at16 b20_60 first 1540.4 warm p50 1171.3 ab-c b1_3 first 223.5 warm p50 155.7 ab-c b3_8 first 354.6 warm p50 271.1 ab-c b8_20 first 668.4 warm p50 591.3 ab-c b20_60 first 2959.4 warm p50 2101.0 ab-d b1_3 first 217.2 warm p50 149.5 ab-d b3_8 first 286.2 warm p50 263.8 ab-d b8_20 first 618.4 warm p50 564.9 ab-d b20_60 first 2735.7 warm p50 2003.5 ab-cf32 b1_3 first 82.5 warm p50 40.7 ab-cf32 b3_8 first 51.0 warm p50 50.4 ab-cf32 b8_20 first 79.6 warm p50 75.5 ab-cf32 b20_60 first 254.3 warm p50 186.2 ab-cf16 b1_3 first 123.5 warm p50 88.6 ab-cf16 b3_8 first 97.6 warm p50 98.9 ab-cf16 b8_20 first 134.6 warm p50 123.6 ab-cf16 b20_60 first 304.5 warm p50 244.1 ab-af32 b1_3 first 77.4 warm p50 38.1 ab-af32 b3_8 first 47.8 warm p50 46.7 ab-af32 b8_20 first 67.6 warm p50 67.0 ab-af32 b20_60 first 218.7 warm p50 167.0 ab-b32 b1_3 first 58.9 warm p50 22.9 ab-b32 b3_8 first 43.8 warm p50 26.6 ab-b32 b8_20 first 36.8 warm p50 34.2 ab-b32 b20_60 first 409.8 warm p50 76.7 ab-b16 b1_3 first 66.5 warm p50 28.5 ab-b16 b3_8 first 35.4 warm p50 32.4 ab-b16 b8_20 first 46.3 warm p50 38.8 ab-b16 b20_60 first 372.2 warm p50 66.6 ## GPU memory per process (MiB) {"arm": "ab-a1", "pid": 3208587, "cold_s": 48.2, "rest_after_warmup_mib": 932, "peak_utterance_block_mib": 5798, "peak_whole_life_mib": 38566, "last_mib": 38566, "samples": 30058} {"arm": "ab-a2", "pid": 3211852, "cold_s": 47.8, "rest_after_warmup_mib": 922, "peak_utterance_block_mib": 4764, "peak_whole_life_mib": 4764, "last_mib": 4764, "samples": 29036} {"arm": "ab-a50", "pid": 3215242, "cold_s": 47.5, "rest_after_warmup_mib": 922, "peak_utterance_block_mib": 4764, "peak_whole_life_mib": 4764, "last_mib": 4764, "samples": 22302} {"arm": "ab-apl", "pid": 3218510, "cold_s": 48.0, "rest_after_warmup_mib": 930, "peak_utterance_block_mib": 4772, "peak_whole_life_mib": 4772, "last_mib": 4772, "samples": 22003} {"arm": "ab-at16", "pid": 3221843, "cold_s": 47.7, "rest_after_warmup_mib": 1698, "peak_utterance_block_mib": 4772, "peak_whole_life_mib": 4772, "last_mib": 4772, "samples": 28137} {"arm": "ab-c", "pid": 3225020, "cold_s": 48.3, "rest_after_warmup_mib": 916, "peak_utterance_block_mib": 4758, "peak_whole_life_mib": 10902, "last_mib": 10902, "samples": 32579} {"arm": "ab-d", "pid": 3228280, "cold_s": 47.8, "rest_after_warmup_mib": 916, "peak_utterance_block_mib": 4758, "peak_whole_life_mib": 8854, "last_mib": 8854, "samples": 32309} {"arm": "ab-cf32", "pid": 3231568, "cold_s": 45.3, "rest_after_warmup_mib": 3874, "peak_utterance_block_mib": 6946, "peak_whole_life_mib": 6946, "last_mib": 6946, "samples": 20979} {"arm": "ab-cf16", "pid": 3234656, "cold_s": 45.6, "rest_after_warmup_mib": 2776, "peak_utterance_block_mib": 4824, "peak_whole_life_mib": 4824, "last_mib": 4824, "samples": 24866} {"arm": "ab-af32", "pid": 3237908, "cold_s": 44.6, "rest_after_warmup_mib": 3932, "peak_utterance_block_mib": 7004, "peak_whole_life_mib": 7004, "last_mib": 7004, "samples": 20764} {"arm": "ab-b32", "pid": 3240981, "cold_s": 12.1, "rest_after_warmup_mib": 3836, "peak_utterance_block_mib": 4800, "peak_whole_life_mib": 5620, "last_mib": 4800, "samples": 19712} {"arm": "ab-b16", "pid": 3242482, "cold_s": 11.8, "rest_after_warmup_mib": 4802, "peak_utterance_block_mib": 5126, "peak_whole_life_mib": 5620, "last_mib": 5126, "samples": 19923}