• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

New NVIDIA GPU roadmap and Pascal details revealed at GTC 2015 opening keynote

Sulik2

Member
Dang I need a new video card this year. Was hoping next gen nvidia would arrive. May have to look at an AMD for the Witcher 3.
 

finalflame

Gold Member
High end nvidia card won't be worth it? I was thinking about upgrading for the witcher 3. I have a gtx 780.

980 by itself won't be enough if you want to max it out with all the bells and whistles + MSAA, even at 1080p60fps, almost guaranteed (I'd love to be proven wrong). Witcher is known for making huge strides as far as graphical requirements go, and I expect Witcher 3 to be no different.

A TitanX, maybe. Also depends on if we get good driver optimizations and how closely the folks at CDPR are working with nVidia.
 
For those who picked up their GTX 970 from Microcenter, I was just contacted by customer support and was informed that they are indeed taking the card back for a refund, not exchange. And I got mine on day one, apparently it's not an issue.

"Hello,

Micro Center at this time is allowing refunds for the card, not exchanges. Please take the card to your local store for the refund.

Thank you for contacting Micro Center. Have a great day!"

So if you're looking at either the GTX 980 or Titan or even another card, Microcenter will take care of it.

There's an open box Sapphire 295X2 for $599.99 at my microcenter, may just take my GTX 970 back and return the other one I just ordered and get that. Or just get a GTX 980 and be content for a while.
 

Velurian

Member
So wait.

If we have - in 2016 - High End cards that are 10x faster than current high end then didnt just...

1. VR entry level go down about 5-6x fiscally (2x980GTX today to power 1440p @ 90 fps vs single mirange 200 € card in a year)

2. 4k and beyond gaming just became affordable


... I mean, in a year, 400 € gaming pc will have A LOT of compute power. How will software scale up with this? Badly optimized games?
 
So what I am I to conclude form this?

If you've paid attention for the past few years, you should take away from this essentially nothing. Every new architecture, every new advance will be talked up endlessly. In this case, as it often is, the big advances won't line up to create a massive GPU performance improvement in the gaming space. Pascal will bring with it moderate gaming performance increases, maybe a little better than Maxwell but probably not by much. If we got as much as a 30% YoY boost between cards at the same naming-level I would be impressed, if we got greater than that I would be fucking gobsmacked.

If we got a 10x increase I would be killing myself to force my body to wake up from this dream state.
 

AmyS

Member
Nvidia Pascal GP100 GPU Flagship Will Pack A 4096-bit Memory Bus And Four 8-Hi HBM2 Stacks

Nvidia’s big Pascal GPU code named GP100 will feature a massive 4096bit bus and four HBM2 stacks each up to 8-Hi. The upcoming Nvidia flagship Pascal chip set to debut on TSMC’s 16nm FinFET process later next year. We have confirmed with our sources that the GPU will be made with two different variations of stacked HBM2 solutions, however both will feature a massive 4096bit memory interface just like AMD’s flagship Fiji GPU launched last month.

The first variation will pack four HBM2 stacks, each will be 4-Hi and will be clocked at 1Ghz. This will go into the traditional consumer GeForce line of GP100 based products. The second variation is also equipped with four HBM2 stacks clocked at 1Ghz, however each will be 8-Hi.
 

AJLma

Member
More memory bandwidth = higher framerates at higher resolutions. 16nm = more cores at higher frequencies. If I'm not mistaken.

I wonder how fast AMD will get the Fury 2 out.
 

paolo11

Member
More memory bandwidth = higher framerates at higher resolutions. 16nm = more cores at higher frequencies. If I'm not mistaken.

I wonder how fast AMD will get the Fury 2 out.


How powerful compared to maxwell (say a single 980gtd )?
 

AmyS

Member
I don't think we'll see Pascal availability until the middle of 2016 at the earliest. Probably fall 2016.
 

AJLma

Member
How powerful compared to maxwell (say a single 980gtd )?

No telling that without knowing the specs of the card.

The Fury X sort of proved that memory bandwidth hasn't been the bottleneck with most games released until today, so 16nm is probably going to be where the real performance gains are achieved, with HBM just providing a higher ceiling.

My opinion, if nVidia is launching with their big card first, it's likely going to outright dethrone the Titan X instead of being a power/performance showpiece like the 970, so I'm expecting a pretty big improvement. nVidia has thrown out huge numbers in the past about theoretical speeds of Pascal, so it should be good.
 
Why did they hold off on Maxwell so long only to give it such a small window? Maxwell only got the 900 series.
Maxwell got the 750 (Ti), 800 series on laptops, and 900 series on desktops/laptops, so it's not like they got nothing out of the architecture.

Hell, if anything, it saves them from another round of "it's on 28nm again?" bitching.
 

BashNasty

Member
I don't think we'll see Pascal availability until the middle of 2016 at the earliest. Probably fall 2016.

Boy, I really hope it's earlier than that. I'm chomping at the bit to put together a brand new computer, and spring is as late as I would want to wait. I'm sitting with a 970 right now though, so I'm not gonn build anything new before pascal.
 

charsace

Member
Maxwell got the 750 (Ti), 800 series on laptops, and 900 series on desktops/laptops, so it's not like they got nothing out of the architecture.

Hell, if anything, it saves them from another round of "it's on 28nm again?" bitching.

Only some 860m chips were Maxwell. The rest were Kepler. So all they got out it was the 750, some versions of the 860m and the 900 series. Kepler was 600-800 series. Fermi has been 400 to 800 series. I wonder why Maxwell didn't come until the middle of 2014.
 
Top Bottom