You are viewing a single comment's thread from:
RE: My quest to use an NVIDIA TITAN V for BOINC (TITAN V & 1080 Ti BOINC Benchmarks)
Great research!
While on paper they have almost the same single precision FLOPS count, Titan V can do even 3 times better in real world applications.
Due to purchase price difference, it seems only Collatz and Milyway are projects where running Titan V is cost optimal vs running GTX 1080 Ti and cost might be close at PrimeGrid.
Are you going to test them on Amicable and Einstein maybe?
I believe it might be down to a difference in the way GV100 is designed at a fundamental level. To quote this anandtech article:
So this seems to make it perform better with integer-type (Primegrid/Collatz?) compute applications. Anyway that's just my speculation.
Yeah pretty much thanks to the price tag, I'm gunna keep it crunching more earnings-focused projects for now. Although I prefer crunching Primegrid over Collatz personally since I like Primegrid's objective more. I'll write a followup article with some more benchmarks, I'll be sure to include Amicable and Einstein.