GTX 1080 CUDA performance worse than 980 Ti
You are currently reading a thread in /g/ - Technology
kek
so much for "muh machine learning"
>>54602907
And this is a surprise?
Anything under the Ti by Nvidia is a meme
>>54602907
Literally where have you seen machine learning associated with the 1080? It's a gaming card and is being marketed as such. The GP100 aka Big Pascal with HBM2 that you likely won't see till next year is where Nvidia is focusing for the workstation market.
>>54603007
>he doesn't know researchers use GeForce for most workloads because Teslas and Quadros are ridiculously expensive
>>54603140
>he doesn't know that real actual researchers use subsidized hardware provided for them by their institution.
Sure, buy whatever suits your basement-dweller needs.
>>54603242
>he doesn't know there are small scale systems running in small offices to do incremental development before running the completed code on massive Tesla clusters
>he doesn't know those small systems are usually GeForces
>>54603284
Cyka blat
>>54603354
Fuck to you sun of bitch make this comment in my face ??
>>54602907
I'd expect that, given it's the GP104 architecture, and the 980 Ti was the GK100 architecture. It's the Gx100 chips that are the computer cards, while everything else is focused on gaming.
