1
edit
Changes
→Assignment
What we found was that all the parallelization methods were all very similar. We also tested this with cuda and found cuda to be the fastest.
[[Image:Graph|thumb|widthpx|http://i.imgur.com/GeyIa97.png|640px]]
[[Image:Data|thumb|widthpx|http://i.imgur.com/TP4107j.png|640px]]