NVIDIA Unveils Pascal GPU: 16GB of memory, 1TB/s Bandwidth

I hope this is 2.5X times speed of a 980 TI. I will buy it instantly and would pay up to 1000 bucks for it.
 
Running 4.6 WC and still waiting for a game that utilizes that. With all the shit unoptimized POS4 and XBONE ports, I don't see much need for a high OC unless benchmarking for ePeen

Please let me know if there exists a game right now where even 4.5 is needed. The only reason I'm even running that is for nongaming purposes

You clearly don't play ArmA.
 
Lol actually no, I don't, because the internet latency here sucks. Haven't been playing many FPSers at all because of it.
 
Zarathustra[H];1041979988 said:
16GB seems excessive considering nothing comes close to my current 6GB cards, but this is possibly the figure for the Tesla cards with consumer models having more reasonable numbers.

Yeah, that's likely just the max it can address, and maybe will show up in their Titan, but the "regular" cards ($600 and down) I reckon will still be 4-6.
 
Isn't HBM in multiples of 4? With HBM2 starting at 8GB?

Cards these powerful should have 8GB+ IMO.
 
I think I understand the point you're trying to make, but then how would you go about comparing game optimization then, especially if they run on different engines as well?

Or is your point terms such as "poorly optimized" inherently meaningless because each game is different?

Certainly optimization, or really lack of is what people are concerned with, is an real issue. It's certainly fair to say that AC:Unity was less optimized at launch compared to it's current iteration as an example, as the results speak for themselves. In this case one would for sure have to question the games optimization.

But it's getting thrown around way too much and vaguely without people actually considering everything involved. Games are really variable kind of like an apples to oranges situation. This is why I find it very doubtful regarding how people can make some definitive comments based on how one game is better optimized seemingly so easily (as they throw it out in commentary).

What's related to this and worse is how it's becoming common now for people to criticize game engines along the lines of Game A would be better if they used the engine from Game B based on how Game A and B look (and by this I mean via screenshots typically, which is another thing I feel people should reconsider). The problem is this is a gross oversimplification of everything involved.

Keep in mind I'm not really discouraging discussion but my issue is more with bold and definitive statements regarding these issues.

Isn't HBM in multiples of 4? With HBM2 starting at 8GB?

Cards these powerful should have 8GB+ IMO.

16GB is a multiple of 4.
 
Sorry, I was addressing the people saying 4-6GB. I should have quoted. IIRC HBM2 starts at 8GB... so I'd expect at least 8GB in mid range +.

8GB utilizing 4 stacks, or 2GB/stack. 4GB and 6GB are definitely possible, but not sure if the process is cheap enough for such cards.
 
You clearly don't play ArmA.

Red Orchestra 2 also has high CPU load on 64 player servers due to all the bullet physics.

It's not that you can't play it on lower end CPU's. You can.

You just won't get the frame rates that I expect, dropping below 60fps in high intensity time periods.
 
Hopefully there'll be 4 GB for the A60 GPUs, 8 GB for the A70 GPUs, 16 GB for the A80 GPUs, and 32 GB for AX and Tesla GPUs.
 
I think I understand the point you're trying to make, but then how would you go about comparing game optimization then, especially if they run on different engines as well?

Or is your point terms such as "poorly optimized" inherently meaningless because each game is different?

That'd be my point. The people claiming s/w is poorly optimized rarely have the information to make that determination. You can claim a game performs poorly, but as a rule, that's all the layman can say.

The worst software I've worked with is enterprise s/w. Specifically, the front end tools for back end systems. They're consistently bad. One example is BRM (Oracle). It's front end is slow (though our current front end is even worse). Actually, BRM's back end wasn't very good either.
 
That still a rather superficial and open (poorly defined) metric.

Yes it is. A system can only do so much in a given time frame. Your machine can only handle so many polygons, textures, particle effects and objects Developers can essentially spend that performance budget however they choose. Rough visual equality between games is just that. MMO's tend to run worse and look worse than FPS games because those resources are used differently. People need to understand that games made with different engines even when they appear somewhat similar are like comparing apples to oranges. Arkham Knight vs. MGSV is one example tossed around despite the latter taking place in far less complex and dynamic environments than the former.
 
yeah you can't really quantify "optimized or not" between different games, engines etc. Each app has their own limitations due to the way the frame work is written. It would be nice to get the max theoretical out of hardware, but there are always stumbling blocks as the GPU's ALU's are "shared" resources.
 
Zarathustra[H];1041981836 said:
Yeah, I'll pass on the whole VR thing, it's just another fad like 3d TV and smello-o-vision.


Well until we get a Halodeck!
 
Zarathustra[H];1041981836 said:
Yeah, I'll pass on the whole VR thing, it's just another fad like 3d TV and smello-o-vision.

We'll see. I'm not sure that gaming will be the first place that it takes hold, but I think VR is going to happen and become successful. That said, I'm not rushing out to buy an Occulus.
 
Zarathustra[H];1041981836 said:
Yeah, I'll pass on the whole VR thing, it's just another fad like 3d TV and smello-o-vision.

It'll be fun to revisit naysayer quotes like this middle of next year :)
 
The only game I can honestly say I wanna play in VR is Alien: Isolation.

Though, when (if) VR specific/focused titles start coming out that could change.
 
EVERY FLIGHT SIM!
EVERY DRIVING SIM!
EVERY FPS!

other then that I think VR is just a money grab and a giant fad.
Totally a super gay fad.
 
HL3? Bwahaha! You'll see the dodobird in the sky before you see HL3. Silly kid, valve delivers games now. They don't make them (unless you count those stupid zombie left for dead 'games' as games). And they won't sell the rights to someone who can and who will.

One of my friends knows one of the people who've done voiceover work for some NPCs on Half Life 1 and 2 (and I think on Episode 1 & 2). I asked him to ask his friend about the status of his work, if any, on Half Life 3, and when he got back to me he said his friend told him that he wasn't at liberty to say due to a non-disclosure agreement, which indicates to me that at least the game is being developed. I look forward to its eventual release.
 
One of my friends knows one of the people who've done voiceover work for some NPCs on Half Life 1 and 2 (and I think on Episode 1 & 2). I asked him to ask his friend about the status of his work, if any, on Half Life 3, and when he got back to me he said his friend told him that he wasn't at liberty to say due to a non-disclosure agreement, which indicates to me that at least the game is being developed. I look forward to its eventual release.

NDA doesn't necessary means it's in development. It must have been at one time long ago and the NDA still stands until now is a big possibility.

There's basically 0 reward and 100% risk for Valve to do HL3 or even Episode 3. Too much hype and expectations. It's probably gonna end up like Duke Nukem.
 
I feel like when this comes out, I will have a strong itch to upgrade from my R9-290.
 
GearVR sold out in less then a day...
There's a big difference between strapping Angry Birds to your face and legitimate VR gaming though.

Just because Samsung Galaxy users are willing to pay $100 to place their Facebook pages two inches in front of their eyes doesn't mean gamers are going to plunk down lots o' cash on the necessary multi-GPU rigs needed to drive the high refresh rate, high resolution headsets that Oculus and Valve are pushing, plus the cost of the headset itself.
 
It seems like something that if done right could be an interesting experience. Not sure I would want to play every game like that, plus I have vertigo issues now so it probably wouldn't work out anyway.

But there are games where immersion is somewhat important and others where it's not really important at all, so I think it depends on the game.
 
Back
Top