Hello,
I have a dataset of about 300 subjects with dwi scans and would like to get a gpu to run bedpostx. From reading the 2013 paper on implementing gpus for bedpostx, I saw that the m2090/c2050 can process 30 slices in < 20 minutes. Since these gpus are going for only about $100 on ebay would purchasing maybe 2 of these be the way to go versus buying one titan (not z or x). Is there a considerable speed-up with the titan?
Thank you for reading,
J
|