Plonky msm的改进版Yao算法

1. 引言

前序博客有：

• Scalar-multiplication算法集
• multiScalar_multiplication Pippeneger算法
• Monero bulletproof 源码结构解析
• https://blog.csdn.net/mutourend/article/details/93752389
• Plonky代码解析

参考书本有《Handbook of Elliptic and Hyperelliptic Curve Cryptography》。

针对：

• https://github.com/mir-protocol/plonky

运行bin/msms.rs文件，有：

  Running `target/debug/msms`
MSMs with terms=2^8, threads=1, window_size=2: avg=0.4473s, avg*threads=0.4473s
MSMs with terms=2^8, threads=2, window_size=2: avg=0.4298s, avg*threads=0.8595s
MSMs with terms=2^8, threads=3, window_size=2: avg=0.4529s, avg*threads=1.3588s
MSMs with terms=2^8, threads=4, window_size=2: avg=0.4227s, avg*threads=1.6908s
MSMs with terms=2^8, threads=5, window_size=2: avg=0.4158s, avg*threads=2.0792s
MSMs with terms=2^8, threads=6, window_size=2: avg=0.4326s, avg*threads=2.5956s
MSMs with terms=2^8, threads=7, window_size=2: avg=0.4310s, avg*threads=3.0171s
MSMs with terms=2^8, threads=8, window_size=2: avg=0.4392s, avg*threads=3.5133s
MSMs with terms=2^8, threads=9, window_size=2: avg=0.4830s, avg*threads=4.3474s
MSMs with terms=2^8, threads=10, window_size=2: avg=0.4875s, avg*threads=4.8747s
MSMs with terms=2^8, threads=11, window_size=2: avg=0.4713s, avg*threads=5.1848s
MSMs with terms=2^8, threads=12, window_size=2: avg=0.4279s, avg*threads=5.1350s
MSMs with terms=2^8, threads=13, window_size=2: avg=0.4431s, avg*threads=5.7597s
MSMs with terms=2^8, threads=14, window_size=2: avg=0.4330s, avg*threads=6.0625s
MSMs with terms=2^8, threads=15, window_size=2: avg=0.4554s, avg*threads=6.8307s
MSMs with terms=2^8, threads=16, window_size=2: avg=0.4297s, avg*threads=6.8752s
MSMs with terms=2^8, threads=17, window_size=2: avg=0.4376s, avg*threads=7.4390s
MSMs with terms=2^8, threads=18, window_size=2: avg=0.4378s, avg*threads=7.8796s
MSMs with terms=2^8, threads=19, window_size=2: avg=0.4897s, avg*threads=9.3041s
MSMs with terms=2^8, threads=20, window_size=2: avg=0.4389s, avg*threads=8.7780s
MSMs with terms=2^8, threads=21, window_size=2: avg=0.4561s, avg*threads=9.5781s
MSMs with terms=2^8, threads=22, window_size=2: avg=0.4286s, avg*threads=9.4285s
MSMs with terms=2^8, threads=23, window_size=2: avg=0.4189s, avg*threads=9.6350s
MSMs with terms=2^8, threads=24, window_size=2: avg=0.4346s, avg*threads=10.4301s
MSMs with terms=2^8, threads=25, window_size=2: avg=0.4224s, avg*threads=10.5588s
MSMs with terms=2^8, threads=26, window_size=2: avg=0.4170s, avg*threads=10.8412s
MSMs with terms=2^8, threads=27, window_size=2: avg=0.4317s, avg*threads=11.6546s
MSMs with terms=2^8, threads=28, window_size=2: avg=0.4287s, avg*threads=12.0023s
MSMs with terms=2^8, threads=29, window_size=2: avg=0.4190s, avg*threads=12.1500s
MSMs with terms=2^8, threads=30, window_size=2: avg=0.4200s, avg*threads=12.6007s
MSMs with terms=2^8, threads=31, window_size=2: avg=0.4264s, avg*threads=13.2184s
MSMs with terms=2^8, threads=32, window_size=2: avg=0.4283s, avg*threads=13.7071s
MSMs with terms=2^8, threads=33, window_size=2: avg=0.4524s, avg*threads=14.9295s
MSMs with terms=2^8, threads=34, window_size=2: avg=0.4358s, avg*threads=14.8186s
MSMs with terms=2^8, threads=35, window_size=2: avg=0.4177s, avg*threads=14.6198s
MSMs with terms=2^8, threads=36, window_size=2: avg=0.4176s, avg*threads=15.0349s
MSMs with terms=2^8, threads=37, window_size=2: avg=0.4199s, avg*threads=15.5355s
MSMs with terms=2^8, threads=38, window_size=2: avg=0.4169s, avg*threads=15.8412s
MSMs with terms=2^8, threads=39, window_size=2: avg=0.4283s, avg*threads=16.7018s
MSMs with terms=2^8, threads=40, window_size=2: avg=0.4430s, avg*threads=17.7187s
MSMs with terms=2^8, threads=41, window_size=2: avg=0.4167s, avg*threads=17.0851s
MSMs with terms=2^8, threads=42, window_size=2: avg=0.4159s, avg*threads=17.4687s
MSMs with terms=2^8, threads=43, window_size=2: avg=0.4177s, avg*threads=17.9600s
MSMs with terms=2^8, threads=44, window_size=2: avg=0.4215s, avg*threads=18.5438s
MSMs with terms=2^8, threads=45, window_size=2: avg=0.4167s, avg*threads=18.7503s
MSMs with terms=2^8, threads=46, window_size=2: avg=0.4168s, avg*threads=19.1714s
MSMs with terms=2^8, threads=47, window_size=2: avg=0.4164s, avg*threads=19.5711s
MSMs with terms=2^8, threads=48, window_size=2: avg=0.4196s, avg*threads=20.1388s
MSMs with terms=2^8, threads=49, window_size=2: avg=0.4165s, avg*threads=20.4072s
MSMs with terms=2^8, threads=50, window_size=2: avg=0.4173s, avg*threads=20.8672s
MSMs with terms=2^8, threads=1, window_size=3: avg=0.2818s, avg*threads=0.2818s
MSMs with terms=2^8, threads=2, window_size=3: avg=0.2808s, avg*threads=0.5616s
MSMs with terms=2^8, threads=3, window_size=3: avg=0.2824s, avg*threads=0.8472s
MSMs with terms=2^8, threads=4, window_size=3: avg=0.2830s, avg*threads=1.1318s
MSMs with terms=2^8, threads=5, window_size=3: avg=0.2855s, avg*threads=1.4277s
MSMs with terms=2^8, threads=6, window_size=3: avg=0.2813s, avg*threads=1.6881s
MSMs with terms=2^8, threads=7, window_size=3: avg=0.2819s, avg*threads=1.9734s
MSMs with terms=2^8, threads=8, window_size=3: avg=0.2808s, avg*threads=2.2461s
MSMs with terms=2^8, threads=9, window_size=3: avg=0.2820s, avg*threads=2.5377s
MSMs with terms=2^8, threads=10, window_size=3: avg=0.2827s, avg*threads=2.8267s
MSMs with terms=2^8, threads=11, window_size=3: avg=0.2858s, avg*threads=3.1437s
MSMs with terms=2^8, threads=12, window_size=3: avg=0.2815s, avg*threads=3.3781s
MSMs with terms=2^8, threads=13, window_size=3: avg=0.2820s, avg*threads=3.6660s
MSMs with terms=2^8, threads=14, window_size=3: avg=0.2835s, avg*threads=3.9686s
MSMs with terms=2^8, threads=15, window_size=3: avg=0.2821s, avg*threads=4.2316s
MSMs with terms=2^8, threads=16, window_size=3: avg=0.2827s, avg*threads=4.5229s
MSMs with terms=2^8, threads=17, window_size=3: avg=0.2823s, avg*threads=4.7994s
MSMs with terms=2^8, threads=18, window_size=3: avg=0.2864s, avg*threads=5.1554s
MSMs with terms=2^8, threads=19, window_size=3: avg=0.2837s, avg*threads=5.3901s
MSMs with terms=2^8, threads=20, window_size=3: avg=0.2829s, avg*threads=5.6581s
MSMs with terms=2^8, threads=21, window_size=3: avg=0.2845s, avg*threads=5.9755s
MSMs with terms=2^8, threads=22, window_size=3: avg=0.2836s, avg*threads=6.2391s
MSMs with terms=2^8, threads=23, window_size=3: avg=0.2840s, avg*threads=6.5310s
MSMs with terms=2^8, threads=24, window_size=3: avg=0.2832s, avg*threads=6.7958s
MSMs with terms=2^8, threads=25, window_size=3: avg=0.2883s, avg*threads=7.2078s
MSMs with terms=2^8, threads=26, window_size=3: avg=0.2852s, avg*threads=7.4164s
MSMs with terms=2^8, threads=27, window_size=3: avg=0.2864s, avg*threads=7.7325s
MSMs with terms=2^8, threads=28, window_size=3: avg=0.2842s, avg*threads=7.9567s
MSMs with terms=2^8, threads=29, window_size=3: avg=0.2841s, avg*threads=8.2396s
MSMs with terms=2^8, threads=30, window_size=3: avg=0.2848s, avg*threads=8.5427s
MSMs with terms=2^8, threads=31, window_size=3: avg=0.2834s, avg*threads=8.7861s
MSMs with terms=2^8, threads=32, window_size=3: avg=0.2882s, avg*threads=9.2230s
MSMs with terms=2^8, threads=33, window_size=3: avg=0.2856s, avg*threads=9.4241s
MSMs with terms=2^8, threads=34, window_size=3: avg=0.2937s, avg*threads=9.9854s
MSMs with terms=2^8, threads=35, window_size=3: avg=0.2859s, avg*threads=10.0078s
MSMs with terms=2^8, threads=36, window_size=3: avg=0.2836s, avg*threads=10.2084s
MSMs with terms=2^8, threads=37, window_size=3: avg=0.2851s, avg*threads=10.5483s
MSMs with terms=2^8, threads=38, window_size=3: avg=0.2842s, avg*threads=10.8015s
MSMs with terms=2^8, threads=39, window_size=3: avg=0.2878s, avg*threads=11.2259s
MSMs with terms=2^8, threads=40, window_size=3: avg=0.2860s, avg*threads=11.4402s
MSMs with terms=2^8, threads=41, window_size=3: avg=0.2840s, avg*threads=11.6448s
MSMs with terms=2^8, threads=42, window_size=3: avg=0.2847s, avg*threads=11.9581s
MSMs with terms=2^8, threads=43, window_size=3: avg=0.2844s, avg*threads=12.2281s
MSMs with terms=2^8, threads=44, window_size=3: avg=0.2838s, avg*threads=12.4851s
MSMs with terms=2^8, threads=45, window_size=3: avg=0.2861s, avg*threads=12.8738s
MSMs with terms=2^8, threads=46, window_size=3: avg=0.2886s, avg*threads=13.2756s
MSMs with terms=2^8, threads=47, window_size=3: avg=0.2870s, avg*threads=13.4896s
MSMs with terms=2^8, threads=48, window_size=3: avg=0.2846s, avg*threads=13.6615s
MSMs with terms=2^8, threads=49, window_size=3: avg=0.2852s, avg*threads=13.9724s
MSMs with terms=2^8, threads=50, window_size=3: avg=0.2860s, avg*threads=14.2988s
MSMs with terms=2^8, threads=1, window_size=4: avg=0.2353s, avg*threads=0.2353s
MSMs with terms=2^8, threads=2, window_size=4: avg=0.2356s, avg*threads=0.4712s
MSMs with terms=2^8, threads=3, window_size=4: avg=0.2389s, avg*threads=0.7166s
MSMs with terms=2^8, threads=4, window_size=4: avg=0.2363s, avg*threads=0.9450s
MSMs with terms=2^8, threads=5, window_size=4: avg=0.2359s, avg*threads=1.1797s
MSMs with terms=2^8, threads=6, window_size=4: avg=0.2370s, avg*threads=1.4221s
MSMs with terms=2^8, threads=7, window_size=4: avg=0.2366s, avg*threads=1.6565s
MSMs with terms=2^8, threads=8, window_size=4: avg=0.2368s, avg*threads=1.8941s
MSMs with terms=2^8, threads=9, window_size=4: avg=0.2395s, avg*threads=2.1558s
MSMs with terms=2^8, threads=10, window_size=4: avg=0.2393s, avg*threads=2.3928s
MSMs with terms=2^8, threads=11, window_size=4: avg=0.2470s, avg*threads=2.7166s
MSMs with terms=2^8, threads=12, window_size=4: avg=0.2599s, avg*threads=3.1187s
MSMs with terms=2^8, threads=13, window_size=4: avg=0.2577s, avg*threads=3.3500s
MSMs with terms=2^8, threads=14, window_size=4: avg=0.2480s, avg*threads=3.4714s
MSMs with terms=2^8, threads=15, window_size=4: avg=0.2457s, avg*threads=3.6853s
MSMs with terms=2^8, threads=16, window_size=4: avg=0.2544s, avg*threads=4.0705s
MSMs with terms=2^8, threads=17, window_size=4: avg=0.2502s, avg*threads=4.2530s
MSMs with terms=2^8, threads=18, window_size=4: avg=0.2539s, avg*threads=4.5703s
MSMs with terms=2^8, threads=19, window_size=4: avg=0.2489s, avg*threads=4.7287s
MSMs with terms=2^8, threads=20, window_size=4: avg=0.2490s, avg*threads=4.9792s
MSMs with terms=2^8, threads=21, window_size=4: avg=0.2496s, avg*threads=5.2425s
MSMs with terms=2^8, threads=22, window_size=4: avg=0.2530s, avg*threads=5.5657s
MSMs with terms=2^8, threads=23, window_size=4: avg=0.2577s, avg*threads=5.9261s
MSMs with terms=2^8, threads=24, window_size=4: avg=0.2500s, avg*threads=6.0005s
MSMs with terms=2^8, threads=25, window_size=4: avg=0.2602s, avg*threads=6.5058s
MSMs with terms=2^8, threads=26, window_size=4: avg=0.2825s, avg*threads=7.3442s
MSMs with terms=2^8, threads=27, window_size=4: avg=0.2826s, avg*threads=7.6301s
MSMs with terms=2^8, threads=28, window_size=4: avg=0.2577s, avg*threads=7.2159s
MSMs with terms=2^8, threads=29, window_size=4: avg=0.2463s, avg*threads=7.1432s
MSMs with terms=2^8, threads=30, window_size=4: avg=0.2458s, avg*threads=7.3736s
MSMs with terms=2^8, threads=31, window_size=4: avg=0.2711s, avg*threads=8.4038s
MSMs with terms=2^8, threads=32, window_size=4: avg=0.2711s, avg*threads=8.6768s
MSMs with terms=2^8, threads=33, window_size=4: avg=0.2707s, avg*threads=8.9338s
MSMs with terms=2^8, threads=34, window_size=4: avg=0.2429s, avg*threads=8.2570s
MSMs with terms=2^8, threads=35, window_size=4: avg=0.2358s, avg*threads=8.2517s
MSMs with terms=2^8, threads=36, window_size=4: avg=0.2433s, avg*threads=8.7586s
MSMs with terms=2^8, threads=37, window_size=4: avg=0.2388s, avg*threads=8.8342s
MSMs with terms=2^8, threads=38, window_size=4: avg=0.2361s, avg*threads=8.9700s
MSMs with terms=2^8, threads=39, window_size=4: avg=0.2345s, avg*threads=9.1451s
MSMs with terms=2^8, threads=40, window_size=4: avg=0.2353s, avg*threads=9.4111s
MSMs with terms=2^8, threads=41, window_size=4: avg=0.2353s, avg*threads=9.6474s
MSMs with terms=2^8, threads=42, window_size=4: avg=0.2348s, avg*threads=9.8612s
MSMs with terms=2^8, threads=43, window_size=4: avg=0.2356s, avg*threads=10.1312s
MSMs with terms=2^8, threads=44, window_size=4: avg=0.2360s, avg*threads=10.3859s
MSMs with terms=2^8, threads=45, window_size=4: avg=0.2358s, avg*threads=10.6110s
MSMs with terms=2^8, threads=46, window_size=4: avg=0.2369s, avg*threads=10.8951s
MSMs with terms=2^8, threads=47, window_size=4: avg=0.2354s, avg*threads=11.0620s
MSMs with terms=2^8, threads=48, window_size=4: avg=0.2359s, avg*threads=11.3248s
MSMs with terms=2^8, threads=49, window_size=4: avg=0.2375s, avg*threads=11.6362s
MSMs with terms=2^8, threads=50, window_size=4: avg=0.2357s, avg*threads=11.7872s
MSMs with terms=2^8, threads=1, window_size=5: avg=0.1938s, avg*threads=0.1938s
MSMs with terms=2^8, threads=2, window_size=5: avg=0.1947s, avg*threads=0.3894s
MSMs with terms=2^8, threads=3, window_size=5: avg=0.1951s, avg*threads=0.5853s
MSMs with terms=2^8, threads=4, window_size=5: avg=0.1965s, avg*threads=0.7859s
MSMs with terms=2^8, threads=5, window_size=5: avg=0.1942s, avg*threads=0.9711s
MSMs with terms=2^8, threads=6, window_size=5: avg=0.1947s, avg*threads=1.1681s
MSMs with terms=2^8, threads=7, window_size=5: avg=0.1954s, avg*threads=1.3678s
MSMs with terms=2^8, threads=8, window_size=5: avg=0.1942s, avg*threads=1.5540s
MSMs with terms=2^8, threads=9, window_size=5: avg=0.1942s, avg*threads=1.7479s
MSMs with terms=2^8, threads=10, window_size=5: avg=0.1946s, avg*threads=1.9460s
MSMs with terms=2^8, threads=11, window_size=5: avg=0.1946s, avg*threads=2.1406s
MSMs with terms=2^8, threads=12, window_size=5: avg=0.1952s, avg*threads=2.3419s
MSMs with terms=2^8, threads=13, window_size=5: avg=0.1948s, avg*threads=2.5319s
MSMs with terms=2^8, threads=14, window_size=5: avg=0.1959s, avg*threads=2.7432s
MSMs with terms=2^8, threads=15, window_size=5: avg=0.1971s, avg*threads=2.9570s
MSMs with terms=2^8, threads=16, window_size=5: avg=0.1964s, avg*threads=3.1431s
MSMs with terms=2^8, threads=17, window_size=5: avg=0.1951s, avg*threads=3.3171s
MSMs with terms=2^8, threads=18, window_size=5: avg=0.1952s, avg*threads=3.5133s
MSMs with terms=2^8, threads=19, window_size=5: avg=0.1955s, avg*threads=3.7151s
MSMs with terms=2^8, threads=20, window_size=5: avg=0.1957s, avg*threads=3.9142s
MSMs with terms=2^8, threads=21, window_size=5: avg=0.1962s, avg*threads=4.1195s
MSMs with terms=2^8, threads=22, window_size=5: avg=0.1957s, avg*threads=4.3064s
MSMs with terms=2^8, threads=23, window_size=5: avg=0.1960s, avg*threads=4.5074s
MSMs with terms=2^8, threads=24, window_size=5: avg=0.1983s, avg*threads=4.7600s
MSMs with terms=2^8, threads=25, window_size=5: avg=0.1980s, avg*threads=4.9502s
MSMs with terms=2^8, threads=26, window_size=5: avg=0.1959s, avg*threads=5.0943s
MSMs with terms=2^8, threads=27, window_size=5: avg=0.1962s, avg*threads=5.2965s
MSMs with terms=2^8, threads=28, window_size=5: avg=0.1966s, avg*threads=5.5043s
MSMs with terms=2^8, threads=29, window_size=5: avg=0.1972s, avg*threads=5.7194s
MSMs with terms=2^8, threads=30, window_size=5: avg=0.1973s, avg*threads=5.9182s
MSMs with terms=2^8, threads=31, window_size=5: avg=0.1969s, avg*threads=6.1040s
MSMs with terms=2^8, threads=32, window_size=5: avg=0.1981s, avg*threads=6.3401s
MSMs with terms=2^8, threads=33, window_size=5: avg=0.1974s, avg*threads=6.5126s
MSMs with terms=2^8, threads=34, window_size=5: avg=0.1976s, avg*threads=6.7191s
MSMs with terms=2^8, threads=35, window_size=5: avg=0.1977s, avg*threads=6.9191s
MSMs with terms=2^8, threads=36, window_size=5: avg=0.1973s, avg*threads=7.1028s
MSMs with terms=2^8, threads=37, window_size=5: avg=0.1983s, avg*threads=7.3380s
MSMs with terms=2^8, threads=38, window_size=5: avg=0.1979s, avg*threads=7.5198s
MSMs with terms=2^8, threads=39, window_size=5: avg=0.1993s, avg*threads=7.7711s
MSMs with terms=2^8, threads=40, window_size=5: avg=0.2003s, avg*threads=8.0110s
MSMs with terms=2^8, threads=41, window_size=5: avg=0.1991s, avg*threads=8.1611s
MSMs with terms=2^8, threads=42, window_size=5: avg=0.1995s, avg*threads=8.3779s
MSMs with terms=2^8, threads=43, window_size=5: avg=0.1993s, avg*threads=8.5711s
MSMs with terms=2^8, threads=44, window_size=5: avg=0.1994s, avg*threads=8.7717s
MSMs with terms=2^8, threads=45, window_size=5: avg=0.2008s, avg*threads=9.0355s
MSMs with terms=2^8, threads=46, window_size=5: avg=0.1994s, avg*threads=9.1734s
MSMs with terms=2^8, threads=47, window_size=5: avg=0.2004s, avg*threads=9.4186s
MSMs with terms=2^8, threads=48, window_size=5: avg=0.2000s, avg*threads=9.6015s
MSMs with terms=2^8, threads=49, window_size=5: avg=0.2009s, avg*threads=9.8461s
MSMs with terms=2^8, threads=50, window_size=5: avg=0.2027s, avg*threads=10.1353s
MSMs with terms=2^8, threads=1, window_size=6: avg=0.1923s, avg*threads=0.1923s
MSMs with terms=2^8, threads=2, window_size=6: avg=0.1947s, avg*threads=0.3895s
MSMs with terms=2^8, threads=3, window_size=6: avg=0.1882s, avg*threads=0.5647s
MSMs with terms=2^8, threads=4, window_size=6: avg=0.1850s, avg*threads=0.7401s
MSMs with terms=2^8, threads=5, window_size=6: avg=0.1841s, avg*threads=0.9203s
MSMs with terms=2^8, threads=6, window_size=6: avg=0.1948s, avg*threads=1.1686s
MSMs with terms=2^8, threads=7, window_size=6: avg=0.1881s, avg*threads=1.3167s
MSMs with terms=2^8, threads=8, window_size=6: avg=0.1948s, avg*threads=1.5583s
MSMs with terms=2^8, threads=9, window_size=6: avg=0.1922s, avg*threads=1.7301s
MSMs with terms=2^8, threads=10, window_size=6: avg=0.1915s, avg*threads=1.9147s
MSMs with terms=2^8, threads=11, window_size=6: avg=0.1887s, avg*threads=2.0757s
MSMs with terms=2^8, threads=12, window_size=6: avg=0.1852s, avg*threads=2.2221s
MSMs with terms=2^8, threads=13, window_size=6: avg=0.1893s, avg*threads=2.4615s
MSMs with terms=2^8, threads=14, window_size=6: avg=0.1928s, avg*threads=2.6994s
MSMs with terms=2^8, threads=15, window_size=6: avg=0.1856s, avg*threads=2.7847s
MSMs with terms=2^8, threads=16, window_size=6: avg=0.1918s, avg*threads=3.0687s
MSMs with terms=2^8, threads=17, window_size=6: avg=0.1889s, avg*threads=3.2120s
MSMs with terms=2^8, threads=18, window_size=6: avg=0.1854s, avg*threads=3.3378s
MSMs with terms=2^8, threads=19, window_size=6: avg=0.1938s, avg*threads=3.6817s
MSMs with terms=2^8, threads=20, window_size=6: avg=0.1911s, avg*threads=3.8222s
MSMs with terms=2^8, threads=21, window_size=6: avg=0.1829s, avg*threads=3.8412s
MSMs with terms=2^8, threads=22, window_size=6: avg=0.1929s, avg*threads=4.2447s
MSMs with terms=2^8, threads=23, window_size=6: avg=0.1932s, avg*threads=4.4430s
MSMs with terms=2^8, threads=24, window_size=6: avg=0.1903s, avg*threads=4.5674s
MSMs with terms=2^8, threads=25, window_size=6: avg=0.1907s, avg*threads=4.7677s
MSMs with terms=2^8, threads=26, window_size=6: avg=0.1934s, avg*threads=5.0275s
MSMs with terms=2^8, threads=27, window_size=6: avg=0.1938s, avg*threads=5.2320s
MSMs with terms=2^8, threads=28, window_size=6: avg=0.1939s, avg*threads=5.4279s
MSMs with terms=2^8, threads=29, window_size=6: avg=0.1969s, avg*threads=5.7101s
MSMs with terms=2^8, threads=30, window_size=6: avg=0.1945s, avg*threads=5.8354s
MSMs with terms=2^8, threads=31, window_size=6: avg=0.1878s, avg*threads=5.8206s
MSMs with terms=2^8, threads=32, window_size=6: avg=0.1913s, avg*threads=6.1205s
MSMs with terms=2^8, threads=33, window_size=6: avg=0.1914s, avg*threads=6.3166s
MSMs with terms=2^8, threads=34, window_size=6: avg=0.1889s, avg*threads=6.4229s
MSMs with terms=2^8, threads=35, window_size=6: avg=0.1959s, avg*threads=6.8559s
MSMs with terms=2^8, threads=36, window_size=6: avg=0.2019s, avg*threads=7.2685s
MSMs with terms=2^8, threads=37, window_size=6: avg=0.1883s, avg*threads=6.9671s
MSMs with terms=2^8, threads=38, window_size=6: avg=0.1855s, avg*threads=7.0507s
MSMs with terms=2^8, threads=39, window_size=6: avg=0.1921s, avg*threads=7.4913s
MSMs with terms=2^8, threads=40, window_size=6: avg=0.1965s, avg*threads=7.8608s
MSMs with terms=2^8, threads=41, window_size=6: avg=0.1960s, avg*threads=8.0357s
MSMs with terms=2^8, threads=42, window_size=6: avg=0.1968s, avg*threads=8.2643s
MSMs with terms=2^8, threads=43, window_size=6: avg=0.1932s, avg*threads=8.3087s
MSMs with terms=2^8, threads=44, window_size=6: avg=0.1968s, avg*threads=8.6603s
MSMs with terms=2^8, threads=45, window_size=6: avg=0.1947s, avg*threads=8.7623s
MSMs with terms=2^8, threads=46, window_size=6: avg=0.1915s, avg*threads=8.8087s
MSMs with terms=2^8, threads=47, window_size=6: avg=0.2000s, avg*threads=9.3981s
MSMs with terms=2^8, threads=48, window_size=6: avg=0.1940s, avg*threads=9.3127s
MSMs with terms=2^8, threads=49, window_size=6: avg=0.1984s, avg*threads=9.7198s
MSMs with terms=2^8, threads=50, window_size=6: avg=0.1990s, avg*threads=9.9486s
MSMs with terms=2^8, threads=1, window_size=7: avg=0.2171s, avg*threads=0.2171s
MSMs with terms=2^8, threads=2, window_size=7: avg=0.1592s, avg*threads=0.3184s
MSMs with terms=2^8, threads=3, window_size=7: avg=0.1625s, avg*threads=0.4876s
MSMs with terms=2^8, threads=4, window_size=7: avg=0.1618s, avg*threads=0.6471s
MSMs with terms=2^8, threads=5, window_size=7: avg=0.1611s, avg*threads=0.8057s
MSMs with terms=2^8, threads=6, window_size=7: avg=0.1601s, avg*threads=0.9604s
MSMs with terms=2^8, threads=7, window_size=7: avg=0.1627s, avg*threads=1.1386s
MSMs with terms=2^8, threads=8, window_size=7: avg=0.1631s, avg*threads=1.3048s
MSMs with terms=2^8, threads=9, window_size=7: avg=0.1604s, avg*threads=1.4433s
MSMs with terms=2^8, threads=10, window_size=7: avg=0.1598s, avg*threads=1.5981s
MSMs with terms=2^8, threads=11, window_size=7: avg=0.1648s, avg*threads=1.8123s
MSMs with terms=2^8, threads=12, window_size=7: avg=0.1640s, avg*threads=1.9682s
MSMs with terms=2^8, threads=13, window_size=7: avg=0.1633s, avg*threads=2.1223s
MSMs with terms=2^8, threads=14, window_size=7: avg=0.1601s, avg*threads=2.2410s
MSMs with terms=2^8, threads=15, window_size=7: avg=0.1628s, avg*threads=2.4418s
MSMs with terms=2^8, threads=16, window_size=7: avg=0.1642s, avg*threads=2.6277s
MSMs with terms=2^8, threads=17, window_size=7: avg=0.1638s, avg*threads=2.7842s
MSMs with terms=2^8, threads=18, window_size=7: avg=0.1653s, avg*threads=2.9753s
MSMs with terms=2^8, threads=19, window_size=7: avg=0.1639s, avg*threads=3.1140s
MSMs with terms=2^8, threads=20, window_size=7: avg=0.1582s, avg*threads=3.1630s
MSMs with terms=2^8, threads=21, window_size=7: avg=0.1626s, avg*threads=3.4153s
MSMs with terms=2^8, threads=22, window_size=7: avg=0.1653s, avg*threads=3.6365s
MSMs with terms=2^8, threads=23, window_size=7: avg=0.1652s, avg*threads=3.7996s
MSMs with terms=2^8, threads=24, window_size=7: avg=0.1627s, avg*threads=3.9049s
MSMs with terms=2^8, threads=25, window_size=7: avg=0.1652s, avg*threads=4.1312s
MSMs with terms=2^8, threads=26, window_size=7: avg=0.1659s, avg*threads=4.3137s
MSMs with terms=2^8, threads=27, window_size=7: avg=0.1669s, avg*threads=4.5052s
MSMs with terms=2^8, threads=28, window_size=7: avg=0.1675s, avg*threads=4.6905s
MSMs with terms=2^8, threads=29, window_size=7: avg=0.1588s, avg*threads=4.6045s
MSMs with terms=2^8, threads=30, window_size=7: avg=0.1687s, avg*threads=5.0610s
MSMs with terms=2^8, threads=31, window_size=7: avg=0.1648s, avg*threads=5.1088s
MSMs with terms=2^8, threads=32, window_size=7: avg=0.1697s, avg*threads=5.4290s
MSMs with terms=2^8, threads=33, window_size=7: avg=0.1642s, avg*threads=5.4191s
MSMs with terms=2^8, threads=34, window_size=7: avg=0.1679s, avg*threads=5.7103s
MSMs with terms=2^8, threads=35, window_size=7: avg=0.1687s, avg*threads=5.9056s
MSMs with terms=2^8, threads=36, window_size=7: avg=0.1686s, avg*threads=6.0687s
MSMs with terms=2^8, threads=37, window_size=7: avg=0.1694s, avg*threads=6.2693s
MSMs with terms=2^8, threads=38, window_size=7: avg=0.1671s, avg*threads=6.3501s
MSMs with terms=2^8, threads=39, window_size=7: avg=0.1634s, avg*threads=6.3709s
MSMs with terms=2^8, threads=40, window_size=7: avg=0.1704s, avg*threads=6.8145s
MSMs with terms=2^8, threads=41, window_size=7: avg=0.1674s, avg*threads=6.8651s
MSMs with terms=2^8, threads=42, window_size=7: avg=0.1734s, avg*threads=7.2837s
MSMs with terms=2^8, threads=43, window_size=7: avg=0.1712s, avg*threads=7.3635s
MSMs with terms=2^8, threads=44, window_size=7: avg=0.1656s, avg*threads=7.2882s
MSMs with terms=2^8, threads=45, window_size=7: avg=0.1684s, avg*threads=7.5797s
MSMs with terms=2^8, threads=46, window_size=7: avg=0.1743s, avg*threads=8.0171s
MSMs with terms=2^8, threads=47, window_size=7: avg=0.1748s, avg*threads=8.2175s
MSMs with terms=2^8, threads=48, window_size=7: avg=0.1674s, avg*threads=8.0342s
MSMs with terms=2^8, threads=49, window_size=7: avg=0.1737s, avg*threads=8.5105s
MSMs with terms=2^8, threads=50, window_size=7: avg=0.1756s, avg*threads=8.7791s
MSMs with terms=2^8, threads=1, window_size=8: avg=0.2881s, avg*threads=0.2881s
MSMs with terms=2^8, threads=2, window_size=8: avg=0.2213s, avg*threads=0.4426s
MSMs with terms=2^8, threads=3, window_size=8: avg=0.2127s, avg*threads=0.6381s
MSMs with terms=2^8, threads=4, window_size=8: avg=0.1865s, avg*threads=0.7458s
MSMs with terms=2^8, threads=5, window_size=8: avg=0.1872s, avg*threads=0.9362s
MSMs with terms=2^8, threads=6, window_size=8: avg=0.1886s, avg*threads=1.1318s
MSMs with terms=2^8, threads=7, window_size=8: avg=0.1943s, avg*threads=1.3602s
MSMs with terms=2^8, threads=8, window_size=8: avg=0.1877s, avg*threads=1.5019s
MSMs with terms=2^8, threads=9, window_size=8: avg=0.1879s, avg*threads=1.6908s
MSMs with terms=2^8, threads=10, window_size=8: avg=0.1884s, avg*threads=1.8844s
MSMs with terms=2^8, threads=11, window_size=8: avg=0.1887s, avg*threads=2.0755s
MSMs with terms=2^8, threads=12, window_size=8: avg=0.1895s, avg*threads=2.2741s
MSMs with terms=2^8, threads=13, window_size=8: avg=0.1899s, avg*threads=2.4689s
MSMs with terms=2^8, threads=14, window_size=8: avg=0.1891s, avg*threads=2.6479s
MSMs with terms=2^8, threads=15, window_size=8: avg=0.1894s, avg*threads=2.8403s
MSMs with terms=2^8, threads=16, window_size=8: avg=0.1906s, avg*threads=3.0493s
MSMs with terms=2^8, threads=17, window_size=8: avg=0.1893s, avg*threads=3.2176s
MSMs with terms=2^8, threads=18, window_size=8: avg=0.1920s, avg*threads=3.4569s
MSMs with terms=2^8, threads=19, window_size=8: avg=0.1909s, avg*threads=3.6273s
MSMs with terms=2^8, threads=20, window_size=8: avg=0.1963s, avg*threads=3.9250s
MSMs with terms=2^8, threads=21, window_size=8: avg=0.2080s, avg*threads=4.3688s
MSMs with terms=2^8, threads=22, window_size=8: avg=0.2088s, avg*threads=4.5947s
MSMs with terms=2^8, threads=23, window_size=8: avg=0.1970s, avg*threads=4.5301s
MSMs with terms=2^8, threads=24, window_size=8: avg=0.1961s, avg*threads=4.7073s
MSMs with terms=2^8, threads=25, window_size=8: avg=0.1994s, avg*threads=4.9859s
MSMs with terms=2^8, threads=26, window_size=8: avg=0.1966s, avg*threads=5.1124s
MSMs with terms=2^8, threads=27, window_size=8: avg=0.2024s, avg*threads=5.4660s
MSMs with terms=2^8, threads=28, window_size=8: avg=0.2156s, avg*threads=6.0363s
MSMs with terms=2^8, threads=29, window_size=8: avg=0.2164s, avg*threads=6.2743s
MSMs with terms=2^8, threads=30, window_size=8: avg=0.2024s, avg*threads=6.0720s
MSMs with terms=2^8, threads=31, window_size=8: avg=0.2199s, avg*threads=6.8166s
MSMs with terms=2^8, threads=32, window_size=8: avg=0.2148s, avg*threads=6.8727s
MSMs with terms=2^8, threads=33, window_size=8: avg=0.2202s, avg*threads=7.2667s
MSMs with terms=2^8, threads=34, window_size=8: avg=0.2104s, avg*threads=7.1544s
MSMs with terms=2^8, threads=35, window_size=8: avg=0.2042s, avg*threads=7.1460s
MSMs with terms=2^8, threads=36, window_size=8: avg=0.2062s, avg*threads=7.4233s
MSMs with terms=2^8, threads=37, window_size=8: avg=0.2124s, avg*threads=7.8579s
MSMs with terms=2^8, threads=38, window_size=8: avg=0.2463s, avg*threads=9.3600s
MSMs with terms=2^8, threads=39, window_size=8: avg=0.2268s, avg*threads=8.8451s
MSMs with terms=2^8, threads=40, window_size=8: avg=0.2149s, avg*threads=8.5957s
MSMs with terms=2^8, threads=41, window_size=8: avg=0.2191s, avg*threads=8.9835s
MSMs with terms=2^8, threads=42, window_size=8: avg=0.2056s, avg*threads=8.6341s
MSMs with terms=2^8, threads=43, window_size=8: avg=0.2019s, avg*threads=8.6836s
MSMs with terms=2^8, threads=44, window_size=8: avg=0.2028s, avg*threads=8.9227s
MSMs with terms=2^8, threads=45, window_size=8: avg=0.2151s, avg*threads=9.6776s
MSMs with terms=2^8, threads=46, window_size=8: avg=0.2008s, avg*threads=9.2378s
MSMs with terms=2^8, threads=47, window_size=8: avg=0.2031s, avg*threads=9.5469s
MSMs with terms=2^8, threads=48, window_size=8: avg=0.2161s, avg*threads=10.3711s
MSMs with terms=2^8, threads=49, window_size=8: avg=0.2013s, avg*threads=9.8642s
MSMs with terms=2^8, threads=50, window_size=8: avg=0.2049s, avg*threads=10.2462s