Hey guys
Should I buy a GTX Titan P or wait for the 1080 ti?
buy 4 rx480s
intel graphics for all of your gay man needs :^)
>>56101230
Poorfag?
No -> Buy
Poorfag?
Yes -> Wait
>>56101384
what if youre not a poor fag just want the better card
still buy the titan or wait?
Do you say Ti like necktie or like T.I. the rapper?
>>56102339
>not sayint Tee dot Ay dot
fucking pleb
>>56102339
I saw tie as in Titanium but was recently told I say it wrong so who knows
>>56101513
Titan will likely outperform it by about 1-5%. I'd imagine the TI (if it even happens, I heard there's going to be a Pascal refresh next year) to be more energy efficient too.
Still an awful value, but if power is what you want then go for it. First single card that can respectably handle high end 4K gaming.
I don't really see the Ti being that much cheaper than the Titan, nor that much slower.
You're probably going to save couple of hundred if you wait and get few % less power.
So if you don't have anything else to spend your money on, go for the Titan.
dont Titans have compute power that regular gtx cards dont?
>>56102459
Only 1%? C'mon they are going to put it exactly inbetween the 1080 and Titan X.
>>56102692
That's pretty much how TI models have always been.
Both the 780ti and 980ti made their titan counterparts look like jokes when it came to value.
>>56102792
What was the 780 Ti equivalent in the 900 series again?
I'm wondering because 980 Ti was basically a 1070 and I'm buying a 1170 when it comes out.
>>56102809
The 780ti probably came closest to the 970. Granted, it wasn't as close as the 980ti vs 1070, but I suppose that's to be expected since Pascal architecture is very similar to Maxwell, whereas Kepler was really different.
>>56102846
Ok so if all goes well the 1170 will be a cheaper 1080 Ti in the future.
>>56102858
Possibly. There's rumors that the 1100 series will be a refresh of the 1000 series since the next generation of Nvidia GPUs (Volta) is expected to land in 2018. The reason for this is the 10nm die shrink Volta is suspected to be on isn't fully matured enough yet.
>>56102605
FP64 and some deep learning shit. It's still not Fermi tier though. That was the last time nvidia had complete domination in Compute
>>56102884
They are already die shrinking from 16 to 14 but I don't know which models for.