With AMD's HSA around the corner, how can/will Primegrid be able to take advantage of this architecture for LLR, GFN, and sieves. We are currently developing an OpenCL app for GFN which may need to be modified for HUMA when Kavari comes out, but should be easier to code. But will LLR find any improvement (question for Mersenne forums I'm sure)?. And I would think sieves should be faster on this APU than any past APU/CPU AMD has put out. All you PG programmers/coders out there, what are your thoughts?
Much more info on HUMA and Kavari is here at PCPerspective:
Largest Primes to Date:
As Double Checker: SR5 109208*5^1816285+1 Dgts-1,269,534
As Initial Finder: SR5 243944*5^1258576-1 Dgts-879,713