1440p Max Settings GPU

[EllieThePurpleFuzzy]

i'm sleepy as heck but i had an odd idea which im 90% sure will be a bad reccomendation but heck.

 

if possible to find (and used) 6GB 780

[othertomperson]

If you've even been on this site for the past month you would know it's more then 3% rather then asking for people to prove it. 

 

As far as numbers go I've been hearing a lot of 192 up to 3.5GB but only 28 from 3.5 to 4GB. Difference may be smaller then this, but it isn't 3%. 

 

If you want to "prove" your own 3% figure, there's plenty of threads on this forum that your proof could go towards.

 

That's the thing. I HAVE these cards. I run two GTX 970s in SLI to power games at 4K at max graphics settings. I KNOW that you are talking bollocks from first hand experience. That's why I told you to prove it, because you are parroting rumour to someone who knows better.

 

 

To each their own I suppose. Anti-aliasing is a thing of the past at 4k, personally. 

 

Do you have a 4K monitor or you're just assuming no anti-aliasing looks terrible?

 

Agreed

 

 

I'd honestly rather have 3.5 GB of VRAM that those 3.5 + 0.5 GB BS.

 

The frustrating this about this is that Nvidia will save themselves the hassle and just do this next time. A 3.5GB limit on this card would be really restricting. You are literally saying you'd rather your FPS just flat out BRICK than lose an unnoticeable amount of performance.

[zappian]

290(x) Crossfire is the best.

[othertomperson]

290 SLI is the best.

 

290 CROSSFIRE doesn't perform as well. 290X CROSSFIRE performs slightly better in some specific circumstances.

[zappian]

290 CROSSFIRE doesn't perform as well. 290X CROSSFIRE performs slightly better in some specific circumstances.

 

We have had this discussion , just accept i made a suggestion to the Op and move on.

 

Good day.

[Accursed Entity]

With a 970, in future AAA titles you will start to hit the 3.5GB VRAM bottleneck and performance will suffer. Get a card with proper 4GB like 290, 290X, or 980.

L. O. L.

[othertomperson]

I didnt run into 3.5gb problems with my card. Even if I SLI and dont get problems with 3.5 how long you think i can run 1440p  with ultra. I dont think long and probably i'll just wait.

 

Just get another 970. Swapping to a 290, 290X or a 980 will cost you a lot and you won't notice the difference. At all. Adding another 970 will make a massive difference.

 

Of all the cards being thrown around in this thread the only one that even out-performs the 970 consistently at 1440p is the 980

[Aniallation]

That's the thing. I HAVE these cards. I run two GTX 970s in SLI to power games at 4K at max graphics settings. I KNOW that you are talking bollocks from first hand experience. That's why I told you to prove it, because you are parroting rumour to someone who knows better.

 

The frustrating this about this is that Nvidia will save themselves the hassle and just do this next time. A 3.5GB limit on this card would be really restricting. You are literally saying you'd rather your FPS just flat out BRICK than lose an unnoticeable amount of performance.

Quotes KamiKatze's post, asks for proof when his post just contained all the first-hand experience that you wanted.

[othertomperson]

Quotes KamiKatze's post, asks for proof when his post just contained all the first-hand experience that you wanted.

 

We must be reading a different post, his first-hand experience is nothing but paranoid conjecture that it's going to "crap out" at some non-descript point sooner or later. That 95% of the time it's perfect. No reference as to what the other 5% even are: every example I've seen brought up by anyone have been games that are seriously broken already and don't run well for anyone, regardless of GPU. It's worth noting: when Dying Light stutters for people with multiple 980s, it's probably not the 970s' mythical vram limit causing your issues.

 

I'm not asking much, just for someone who's actually experiencing a fucking problem to say what it is. That is yet to even happen.

 

Alternatively link to a report on this issue that doesn't either:

 

A) State the performance impact of this as 1-3%

B) Open with words to the effect of "we couldn't replicate any impact of this in our testing at all".