4 ms·
I'm really happy our startup didn't go all in on Tesla (Pascal architecture) yet. These look amazing.
by lowglow 9y ago
I'm really happy our startup didn't go all in on Tesla (Pascal architecture) yet. These look amazing.
- mattnewton 9y agoI feel like every time I buy cards, Nividia announces the successor with absurd improvements.
- dom0 9y agoOTOH improvements in the mainstream segments seem to go slower: Mainstream cards are about twice as fast now as they were five years ago.
- lowglow 9y agoYeah, I just sprung for a Titan Xp -- waiting for it to become obsolete next month.
- mattnewton 9y agoWell, close to already if you are looking at $$/comprable performance, with the 1080ti
- mastazi 9y agoThe Titan Xp (with lowercase p, as opposed to the Titan XP) came out after the 1080 Ti so I'm sure GP took the latter into consideration before making a decision...
- lowglow 9y agoYep. I'm not sure it was worth the extra $$ for the extra specs just yet. We'll see when we SLI it. The issue though is no memory sharing with the GTX/Titan line. If that were the case, I probably just would have sprung for two 1080Tis out the gate. Definitely loving the eight 1080Tis they just fit in here though: http://www.velocitymicro.com/promagix-g480-high-performance-computing.php http://www.velocitymicro.com/promagix-g480-high-performance-...
- deepnotderp 9y agoCuda should run on both, right? Unless you're talking about shader assembly or hardware.