As of RevĀ 16142c1
No Format |
---|
basic.read100 5.91 us read single 100B object with 30B key basic.readBw100 1618.17 MB/s bandwidth reading 100B object with 30B key basic.read1K 76.89 us read single 1KB object with 30B key basic.readBw1K 122137.96 MB/s bandwidth reading 1KB object with 30B key basic.read10K 1110.04 us read single 10KB object with 30B key basic.readBw10K 865914.61 MB/s bandwidth reading 10KB object with 30B key basic.read100K 47.42 us read single 100KB object with 30B key basic.readBw100K 2.0 GB/s bandwidth reading 100KB object with 30B key basic.read1M 421420.28 us read single 1MB object with 30B key basic.readBw1M 2.2 GB/s bandwidth reading 1MB object with 30B key basic.write100 1815.37 us write single 100B object with 30B key basic.writeBw100 56.21 MB/s bandwidth writing 100B object with 30B key basic.write1K 2119.9 us write single 1KB object with 30B key basic.writeBw1K 4348.60 MB/s bandwidth writing 1KB object with 30B key basic.write10K 4138.35 us write single 10KB object with 30B key basic.writeBw10K 231247.17 MB/s bandwidth writing 10KB object with 30B key basic.write100K 237235.93 us write single 100KB object with 30B key basic.writeBw100K 400405.83 MB/s bandwidth writing 100KB object with 30B key basic.write1M 2.2 ms write single 1MB object with 30B key basic.writeBw1M 429436.10 MB/s bandwidth writing 1MB object with 30B key broadcast 206193.03 us broadcast message to 9 slaves netBandwidth 1542.12 GBMB/s many clients reading from different servers netBandwidth.max 16.46 MB/s fastest client netBandwidth.min 1012.84 MB/s slowest client readNotFound 1314.65 us read object that doesn't exist # Gauges impact of asynchronous writes on synchronous writes. # Write two values. The size of the first varies over trials # (its size is given as 'firstObjectSize'). The first write is # either synchronous (if firstWriteIsSync is 1) or asynchronous # (if firstWriteIsSync is 0). The response time of the first # write is given by 'firstWriteLatency'. The second write is # a 100 B object which is always written synchronously (its # response time is given by 'syncWriteLatency' # Both writes use a 30 B key. # Generated by 'clusterperf.py writeAsyncSync' # # firstWriteIsSync firstObjectSize firstWriteLatency(us) syncWriteLatency(us) #---------------------------------------------------------------------------- 0 100 1816.62 1916.68 0 1000 2120.96 1816.78 0 10000 40.61 1917.54 0 100000 245238.83 2119.36 0 1000000 21942228.81 2629.37 1 100 1917.12 1619.43 1 1000 23.02 2120.2 1 10000 4047.7 2122.84 1 100000 242251.61 2928.50 1 1000000 21562215.37 2829.76 # RAMCloud multiWrite performance for 100 B objects with 30 byte keys # located on a single master. # Generated by 'clusterperf.py multiWrite_oneMaster' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 1 1 1 1816.16 1816.16 2 1 2 2120.82 10.91 3 1 3 2422.14 87.05 4 1 4 2625.63 6.63 5 1 5 2926.09 5.84 6 1 6 3027.7 54.16 7 1 7 3430.80 54.03 8 1 8 3432.50 4.30 9 1 9 3633.46 43.07 10 1 10 3937.90 43.07 20 1 20 6155.58 32.18 30 1 30 8273.2 2.74 40 1 40 9389.80 2.32 50 1 50 109107.4 2.21 60 1 60 137123.62 2.31 70 1 70 152143.29 2.21 80 1 80 173168.62 2.21 90 1 90 181.2 2.0 100 1 100 238196.82 2.40 200 1 200 461353.36 21.38 300 1 300 502516.04 1.7 400 1 400 650677.98 1.67 500 1 500 815826.69 1.67 600 1 600 961987.54 1.6 700 1 700 11181142.89 1.6 800 1 800 17371300.90 21.26 900 1 900 19281465.49 21.16 1000 1 1000 21571634.40 21.26 2000 1 2000 32883176.91 1.6 3000 1 3000 47364812.04 1.6 4000 1 4000 62606435.31 1.6 5000 1 5000 78667998.5 1.6 # RAMCloud multiRead performance for 100 B objects with 30 byte keys # located on a single master. # Generated by 'clusterperf.py multiRead_oneMaster' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 1 1 1 5.6.4 5.6.4 2 1 2 87.42 43.26 3 1 3 98.32 32.17 4 1 4 109.75 2.74 5 1 5 1210.25 2.41 6 1 6 1311.09 2.20 7 1 7 1412.27 21.08 8 1 8 1513.29 1.97 9 1 9 1615.26 1.87 10 1 10 16.71 1.76 20 1 20 2523.29 1.32 30 1 30 3332.9 1.1 40 1 40 4239.69 1.10 50 1 50 5250.86 1.10 60 1 60 6159.23 1.0 70 1 70 68.69 1.0 80 1 80 7374.29 0.9 90 1 90 73.85 0.8 100 1 100 7978.35 0.8 200 1 200 126127.54 0.6 300 1 300 175179.58 0.6 400 1 400 219227.61 0.56 500 1 500 270279.08 0.56 600 1 600 329330.21 0.56 700 1 700 389373.85 0.65 800 1 800 447444.81 0.6 900 1 900 487.04 0.5 1000 1 1000 535550.4 0.56 2000 1 2000 10771089.27 0.5 3000 1 3000 15041575.81 0.5 4000 1 4000 20162052.6 0.5 5000 1 5000 25732590.76 0.5 # RAMCloud multiRead performance for 100 B objects with 30 byte keys # with one object located on each master. # Generated by 'clusterperf.py multiRead_oneObjectPerMaster' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 1 1 1 65.5 65.5 2 2 1 76.7 3.93 3 3 1 98.10 32.07 4 4 1 10 9.84 2.74 5 5 1 1110.47 2.31 6 6 1 12.93 2.21 7 7 1 1413.07 2.0 8 8 1 1615.32 21.09 9 9 1 16.58 1.89 10 10 1 1817.90 1.97 11 11 1 3021.29 2.70 12 12 1 3324.09 2.71 13 13 1 3328.28 2.62 14 14 1 2930.57 2.12 15 15 1 3533.18 2.3 16 16 1 4732.51 32.0 17 17 1 3944.98 2.36 18 18 1 39.15 2.2 # RAMCloud multiRead performance for an approximately fixed 19number # of 100 B objects with 30 byte keys # distributed evenly 19across varying number of masters. # Generated by 'clusterperf.py multiRead_general' # # Num Objs 1 Num Masters Objs/Master 40.9 Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 5000 2.2 20 1 20 5000 2563.5 1 43.8 0.5 5000 2.2 2 21 2500 21 3673.8 1 0.7 50.8 4998 2.4 3 22 1666 222662.6 1 0.5 5000 54.6 4 2.5 1250 23 2906.8 23 0.6 1 5000 61.9 5 2.7 1000 24 2682.1 24 0.5 4998 1 72.06 833 3.0 3143.6 25 0.6 25 4998 1 7 58.8 714 4756.2.4 26 1.0 26 5000 1 8 74.4 625 3765.5 2.9 27 0.8 4995 27 9 1 555 66.2 3903.6 2.5 0.8 28 5000 28 10 1 500 654030.19 20.38 4994 29 11 29 454 1 4420.7 67.5 0.9 2.3 4992 30 12 30 416 3914.3 1 73.4 0.8 4992 2.4 13 31 384 31 3934.4 1 0.8 71.2 4998 2.3 14 32 357 32 6303.5 1 1.3 4995 84.0 15 2.6 333 33 3976.7 33 0.8 1 4992 76.8 16 2.3 312 34 4019.9 34 0.8 4998 1 78.4 17 294 2.3 4036.9 35 350.8 4986 1 18 81.9 277 24070.38 36 0.8 # RAMCloud multiRead performance for 36an approximately fixed number # of 100 B objects with 30 byte keys # 1distributed evenly across varying number of masters. # Requests 78.3 2.2 37 37 1 84.8 2.3 38 38 1 87.8 2.3 39 39 1 158.5 4.1 40 40 1 90.7 2.3 41 41 1 93.6 2.3 # RAMCloud multiRead performance for an approximately fixed number # of 100 B objects with 30 byte keys # distributed evenly across varying number of masters. # Generated by 'clusterperf.py multiRead_general' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 5000 1 5000 2545.6 0.5 5000 2 2500 2479.4 0.5 4998 3 1666 2526.9 0.5 5000 4 1250 2694.5 0.5 5000 5 1000 2942.0 0.6 4998 6 833 3075.3 0.6 4998 7 714 3893.2 0.8 5000 8 625 3810.1 0.8 4995 9 555 3765.6 0.8 5000 10 500 3815.8 0.8 4994 11 454 3853.0 0.8 4992 12 416 3859.1 0.8 4992 13 384 4000.7 0.8 4998 14 357 3989.2 0.8 4995 15 333 3938.4 0.8 4992 16 312 4033.4 0.8 4998 17 294 5121.8 1.0 4986 18 277 4788.5 1.0 4997 19 263 4221.2 0.8 5000 20 250 4159.1 0.8 4998 21 238 5287.8 1.1 4994 22 227 4344.4 0.9 4991 23 217 5582.1 1.1 4992 24 208 4269.4 0.9 5000 25 200 4593.6 0.9 4992 26 192 4340.4 0.9 4995 27 185 4412.5 0.9 4984 28 178 4414.7 0.9 4988 29 172 5103.3 1.0 4980 30 166 4444.5 0.9 4991 31 161 5273.3 1.1 4992 32 156 4475.7 0.9 4983 33 151 4593.1 0.9 4998 34 147 4692.6 0.9 4970 35 142 4772.7 1.0 4968 36 138 7377.5 1.5 4995 37 135 4536.2 0.9 4978 38 131 7690.1 1.5 4992 39 128 4653.0 0.9 5000 40 125 6863.8 1.4 4961 41 121 6427.3 1.3 # RAMCloud multiRead performance for an approximately fixed number # of 100 B objects with 30 byte keys # distributed evenly across varying number of masters. # Requests are issued in a random order. # Generated by 'clusterperf.py multiRead_generalRandom' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 5000 1 5000 2779.1 0.6 5000 2 2500 2767.3 0.6 4998 3 1666 2904.5 0.6 5000 4 1250 3119.3 0.6 5000 5 1000 3270.3 0.7 4998 6 833 3453.9 0.7 4998 7 714 3526.2 0.7 5000 8 625 4103.0 0.8 4995 9 555 3980.7 0.8 5000 10 500 3881.7 0.8 4994 11 454 4001.2 0.8 4992 12 416 4223.7 0.8 4992 13 384 4436.8 0.9 4998 14 357 4577.4 0.9 4995 15 333 4782.4 1.0 4992 16 312 4849.1 1.0 4998 17 294 5258.9 1.1 4986 18 277 6063.0 1.2 4997 19 263 6408.4 1.3 5000 20 250 7550.9 1.5 4998 21 238 7818.0 1.6 4994 22 227 7946.1 1.6 4991 23 217 8165.7 are issued in a random order. # Generated by 'clusterperf.py multiRead_generalRandom' # # Num Objs Num Masters Objs/Master Latency (us) Latency/Obj (us) #---------------------------------------------------------------------------- 5000 1.6 1 4992 5000 24 208 8366.32891.6 10.76 5000 25 2 2002500 84262851.05 10.76 49924998 26 3 1921666 88222858.14 10.86 49955000 27 4 1851250 88273033.40 10.86 49845000 28 5 1781000 87073167.17 10.76 4998 4988 296 172833 87583337.81 10.87 49804998 307 166714 88024337.40 10.89 5000 4991 318 161625 89963812.43 10.8 49924995 329 156555 90173798.98 10.8 49835000 3310 151500 91185692.17 1.81 49984994 3411 147454 91653998.76 10.8 49704992 3512 142416 91744140.02 10.8 49684992 3613 138384 93674499.4 10.9 49954998 3714 135357 93934797.42 1.90 49784995 15 38 333 131 126814707.5 20.59 4992 3916 128312 98854799.31 21.0 50004998 17 40 294 125 122495351.29 21.41 49614986 4118 121277 94745860.04 1.92 # Cumulative distribution of time for a single client to read a # single 100-byte object from a single server. Each line indicates # that a given fraction of all reads took at most a given time # to complete. # Generated by 'clusterperf.py readDist' # # Time (usec) Cum. Fraction #--------------------------- 0.00 0.000 54.5970 0.000 54.6878 0.010 54.6979 0.020 54.7079 0.030 54.7080 0.040 54.7080 0.050 54.7081 0.060 54.7181 0.070 54.7182 0.080 54.7182 0.090 54.7183 0.100 54.7183 0.110 54.7283 0.120 54.7284 0.130 54.7284 0.140 54.7284 0.150 54.7284 0.160 54.7285 0.170 54.7285 0.180 54.7385 0.190 54.7385 0.200 54.7385 0.210 54.7386 0.220 54.7386 0.230 54.7386 0.240 54.7386 0.250 54.7386 0.260 54.7486 0.270 54.7487 0.280 54.7487 0.290 54.7487 0.300 54.7487 0.310 54.7487 0.320 54.7487 0.330 54.7488 0.340 54.7588 0.350 54.7588 0.360 54.7588 0.370 54.7588 0.380 54.7588 0.390 54.7588 0.400 54.7588 0.410 54.7588 0.420 54.7689 0.430 54.7689 0.440 54.7689 0.450 54.7689 0.460 54.7689 0.470 54.7689 0.480 54.7789 0.490 54.7789 0.500 54.7789 0.510 54.7790 0.520 54.7790 0.530 54.7890 0.540 54.7890 0.550 54.7890 0.560 54.7890 0.570 54.7890 0.580 54.7990 0.590 54.7990 0.600 54.7990 0.610 54.7991 0.620 54.8091 0.630 54.8091 0.640 54.8091 0.650 54.8091 0.660 54.8191 0.670 54.8192 0.680 54.8192 0.690 54.8192 0.700 54.8293 0.710 54.8293 0.720 54.8394 0.730 54.8495 0.740 54.8596 0.750 54.8697 0.760 54.8798 0.770 54.8999 0.780 5.9100 0.790 5.9301 0.800 5.9503 0.810 5.9605 0.820 5.9707 0.830 65.0112 0.840 65.2247 0.850 65.3952 0.860 65.4154 0.870 65.4356 0.880 65.4457 0.890 65.4759 0.900 65.5065 0.910 65.6291 0.920 6.8403 0.930 6.9407 0.940 6.9610 0.950 6.9812 0.960 76.0215 0.970 76.0829 0.980 7.5626 0.990 1172.2464 0.999 86102.7397 0.9999 338582.0838 1.000 # RAMCloud read performance for 100 B objects # with keys of various lengths. # Generated by 'clusterperf.py readVaryingKeyLength' # # Key Length Latency (us) Bandwidth (MB/s) #---------------------------------------------------------------------------- 1 5.91 0.2 5 5.90 0.89 10 5.91 1.69 15 65.0 2.48 20 54.9 3.29 25 5.90 4.18 30 65.0 45.87 35 65.1 56.56 40 65.02 67.3 45 65.2 68.93 50 65.21 79.74 55 65.31 810.42 60 65.3 910.18 65 65.23 911.97 70 65.25 1012.71 75 65.24 1113.53 80 65.14 1214.52 85 65.14 1315.31 90 65.01 1416.38 95 5.6.1 1416.83 100 65.15 1517.53 200 6.90 2731.78 300 76.16 4043.15 400 7.70 4954.35 500 76.8 6070.90 600 76.9 7282.14 700 86.29 8196.43 800 87.41 90108.40 900 87.52 100119.81 1000 97.04 106129.47 2000 98.23 206230.63 3000 8.9.7 294321.93 4000 109.7 355394.82 5000 1110.52 414469.81 6000 1210.38 466528.43 7000 1311.05 513580.91 8000 1412.02 544623.67 9000 1413.46 595633.71 10000 1513.18 629693.52 20000 2321.19 826871.96 30000 3029.74 932971.48 40000 4039.56 941962.42 50000 5047.09 954995.29 60000 5956.02 9701018.50 # RAMCloud write performance for 100 B objects # with keys of various lengths. # Generated by 'clusterperf.py writeVaryingKeyLength' # # Key Length Latency (us) Bandwidth (MB/s) #---------------------------------------------------------------------------- 1 1715.8 0.1 5 1715.17 0.3 10 1615.8 0.6 15 1715.18 0.89 20 1715.27 1.12 25 1716.1 1.45 30 1715.27 1.78 35 1715.26 2.1.9 40 1715.04 2.25 45 1715.25 2.58 50 1715.48 23.70 55 1816.95 3.2.8 60 1816.43 3.15 65 1816.30 3.49 70 1815.19 34.72 75 1816.70 34.85 80 1816.81 4.17 85 1916.26 4.29 90 1917.51 45.40 95 1917.43 45.72 100 1918.32 45.92 200 2018.35 910.43 300 2019.4 14.08 400 2019.84 1819.47 500 2120.85 2123.92 600 2120.83 2628.32 700 2120.57 3132.12 800 2220.58 3336.97 900 2321.15 3740.10 1000 2322.53 4042.7 2000 2827.25 6769.83 3000 32.65 8788.71 4000 3836.14 100104.08 5000 4342.26 110111.39 6000 4745.09 121124.76 7000 5150.1 130133.73 8000 53.71 142143.26 9000 6257.42 137150.61 10000 6859.02 140161.20 20000 116104.47 163182.82 30000 159145.94 178196.98 40000 209201.02 182189.6 50000 255249.41 186191.74 60000 300293.08 190194.8 # RAMCloud read performance as a function of load (1 or more # clients all reading a single 100-byte object with 30-byte key # repeatedly). # Generated by 'clusterperf.py readLoaded' # # numClients readLatency(us) throughput(total kreads/sec) #---------------------------------------------------------- 1 65.03 166189 2 65.34 320370 3 6.43 469473 4 76.3 547632 5 87.56 588662 6 109.09 603604 7 1110.76 598661 8 1312.2 608656 9 1415.90 605598 10 16.59 607592 11 1817.10 606647 12 1918.86 606645 13 21.39 610595 14 2221.14 633653 15 2325.6 635587 16 2625.35 608629 17 2726.86 611640 18 2928.70 606643 19 3129.2 608650 20 3130.7 631652 # RAMCloud read performance when 1 or more clients read # 100-byte objects with 30-byte keys chosen at random from # 10 servers. # Generated by 'clusterperf.py readRandom' # # numClients throughput(total kreads/sec) slowest(ms) reads > 10us #-------------------------------------------------------------------- 1 160177 0.5759 10.0%7% 2 313324 0.5857 0.5%8% 3 366524 190.9233 0.7%3% 4 620671 0.4564 0.8%6% 5 759741 06.5330 01.9%1% 6 884943 0.6257 1.3%4% 7 10481103 0.6564 1.2%3% 8 11731183 07.8009 1.3%7% 9 12131356 70.4661 1.5%6% 10 10871514 230.0062 21.2%8% 11 15621564 010.5547 1.6%1% 12 16121661 010.5848 21.2%9% 13 18311763 010.6440 1.9%7% 14 18401984 010.5740 21.8%4% 15 11972123 3010.7743 21.1%5% 16 19682130 910.4947 2.2%0% 17 23722281 010.5547 2.1%4% 18 24932337 010.5546 2.7%6% 19 26872726 0.5350 23.4%1% 20 26762857 02.5851 32.1%7% 21 28272749 013.6012 34.3%0% 22 29142452 0.6150 4.1%2% 23 29153225 0.4954 54.0%8% 24 30803182 03.5338 5.2%3% 25 33043366 0.6550 56.0%5% 26 3323 2788 021.5782 56.9%5% 27 33313688 0.5260 76.1%8% 28 35873730 0.5461 7.4%7% 29 36212316 036.5345 810.3%8% 30 36583794 0.4761 812.8%0% 31 37203839 0.5964 913.6%3% 32 36843681 03.5473 1116.0% 33 37173884 2.3812 1215.4%3% 34 37361980 657.6309 13.3%8% 35 41253967 23.5230 1420.8%6% 36 41744028 23.3684 1620.7%3% 37 40132472 228.4643 1722.2%5% 38 42794080 2.5374 1923.3%8% 39 43654281 2.5029 2123.3% 40 4367 3245 162.9031 2124.6%3% 41 39034296 84.7826 2526.6%5% 42 41813989 811.8612 2527.1%9% 43 4387 4225 412.8878 26.9%4% 44 41864452 82.7656 2729.8%3% 45 37184432 312.1960 3032.5%4% 46 35744471 242.3770 3332.2%7% 47 44544530 2.7244 3336.9%0% 48 45094537 42.0146 3436.9%2% 49 4231 4114 114.0072 3538.8%9% 50 2995 4439 30.69 235.967% **** server8.rc11.log: 1376433121.896881152 src/FailureDetector.cc:159 in FailureDetector::pingRandomServer default WARNING[19126:3]: Ping 33.1% timeout to server id 5.0 (locator "infrc:host=192.168.1.105,port=12247") |