Vincent Chen
Vincent Chen
@mclarsen @jameskress @rbinyahib
Hey @mclarsen and @jameskress, I have to agree, there should be some way to manually set the mode/blocksize. Roba was seeing almost perfect scaling with the dynamic scheduler for smaller...
Here is some scaling I tried today on Alaska with the Enzo dataset. These are just times for the sampling phase only. Interactive: https://plot.ly/~applekey/4/enzo-map-sample-scaling/
Sorry, had some previous run result data jumbled in that graph. https://plot.ly/~applekey/7/static-dynamic-guided-auto/
Yeah all the data seems to point towards that. It would be interesting to doctor a dataset where we can change the size, procedurally generate a uniform field of tets,...
Yeah, Roba's manual testing seemed to come to same conclusion. On Fri, Feb 5, 2016 at 1:00 PM, Jeremy Meredith [email protected] wrote: > Am I reading those plots wrong, or...