Y
Hacker News
new
|
ask
|
show
|
jobs
by
lamontcg
926 days ago
Practically does using unified memory mean that the slow transfer of training/testing data to the GPU would be eliminated?