nVidia 600 Series: Release and effect it'll have on 5xx series.

[Evildeffy]

2 3gb GTX 580s would be best for surround gaming. 1GB of ram is good for a 1920x1080 screen, 3gb therefore is good for 3 screens. Trying to put 3 screens worth of resolution into the GTX 570s 1.25gb memory is going to cause performance issues, especially when you crank up AA/AF. Even the 2gb AMD cards would suffer a bit.

If you are putting the money into a triple monitor surround setup then Id suggest going with dual 3gb GTX 580s.

Where in the name of Zeus' butthole did you find the logic of "1GB is good for 1 screen therefor 3GB should be good for 3 screens"?

2GB of VRAM per GPU at this point is good enough for triple monitor gaming, even Metro 2033 can run at full res with triple screens with "only" 2GB of VRAM per GPU.

[Deleted]

Where in the name of Zeus' butthole did you find the logic of "1GB is good for 1 screen therefor 3GB should be good for 3 screens"?

2GB of VRAM per GPU at this point is good enough for triple monitor gaming, even Metro 2033 can run at full res with triple screens with "only" 2GB of VRAM per GPU.

People seem to think that the VRAM stacks? Or?

[BicycleMafioso]

People seem to think that the VRAM stacks? Or?

I think you are talking about VRAM stacking in SLI/Crossfire.

But there are special 3GiB versions of the GTX580

[Deleted]

I think you are talking about VRAM stacking in SLI/Crossfire.

But there are special 3GiB versions of the GTX580

I'm aware. It sounded like 2 posters (one of which I quoted) thought that VRAM stacks in SLI/Crossfire - which it doesn't - and was probably a bit misleading for the OP.

[Evildeffy]

I'm aware. It sounded like 2 posters (one of which I quoted) thought that VRAM stacks in SLI/Crossfire - which it doesn't - and was probably a bit misleading for the OP.

No he's deffo not talking about stacking the VRAM, he's referring to 3GB GTX580s * 2 then going on about 3GB is meant for 1GB per screen...

Which was funny to say the least, also horribly incorrect ofcourse.

[Vallin]

Isn't the 595 supposed to come out by the end of the year? I have my hopes up for that card.

[shoebox]

Can you even have 590 in 4 way SLI? Or at least 3?

yes you can.

[Evildeffy]

yes you can.

No you cannot.

The GTX590 can be SLI'ed once, after which it becomes a QUAD SLI machine due to the fact that the GTX590 is a dual GPU card.

If you refer to the picture earlier, those are compute machines and not SLI linking, Quad SLI/X-Fire is the limit at this point.
Which means you can f.ex link 4 * HD6970 or link 2 * HD6990, or the nVidia equivalent 4 * GTX580 or 2 * GTX590 (SPECIAL MOBOS ONLY, NF200 CHIP REQUIRED!)

[shoebox]

No you cannot.

The GTX590 can be SLI'ed once, after which it becomes a QUAD SLI machine due to the fact that the GTX590 is a dual GPU card.

If you refer to the picture earlier, those are compute machines and not SLI linking, Quad SLI/X-Fire is the limit at this point.
Which means you can f.ex link 4 * HD6970 or link 2 * HD6990, or the nVidia equivalent 4 * GTX580 or 2 * GTX590 (SPECIAL MOBOS ONLY, NF200 CHIP REQUIRED!)

oh my mistake, i thought he meant can you put the GTX 590's in a quad sli configuration (4gpu's being 2 gtx 590's).

[Holyshnikies]

http://news.softpedia.com/news/Nvidi...t-209623.shtml

Nvidia pushing their 600 series to Q1 2012.
AMD releasing their news cards in Q4 2011.

Id rather wait for nvidia... but thats just me....

[Asmekiel]

http://news.softpedia.com/news/Nvidi...t-209623.shtml

Nvidia pushing their 600 series to Q1 2012.
AMD releasing their news cards in Q4 2011.

Id rather wait for nvidia... but thats just me....

Is it just a die-shrink or a totally new architecture? Getting confused...

[FlawlessSoul]

Is it just a die-shrink or a totally new architecture? Getting confused...

Quote from that exact article:

Kepler is the name used by Nvidia for its next-generation graphics core and, just like AMD's Southern Islands GPUs, is a 28nm die shrink of the current Fermi architecture with a few minor tweaks and improvements meant to increase the performance of the chip.

[Asmekiel]

Quote from that exact article:

Yea, I read that. But earlier people said it was a new architecture with the same relative impact of Sandybridge.
That's what got me confused. But I'm going to assume the article is right.

So what's in it for us? A smaller die with less power usage and small performance increase or is it going to be the same diesize which would result in way more performance but not so much less power usage.

-Sorry for that maybe stupid question, but I didn't really keep up with all that stuff lately. Bastion was more interesting

-edit- What tetris says.

[BicycleMafioso]

All I've heard before was; New architecture, and shrink. But not a Fermi-shrink.

[hayharut]

i heard that they were going to be released in the 1st Q of 2012

[FlawlessSoul]

Yea, I read that. But earlier people said it was a new architecture with the same relative impact of Sandybridge.
That's what got me confused. But I'm going to assume the article is right.

So what's in it for us? A smaller die with less power usage and small performance increase or is it going to be the same diesize which would result in way more performance but not so much less power usage.

-Sorry for that maybe stupid question, but I didn't really keep up with all that stuff lately. Bastion was more interesting

-edit- What tetris says.

You're right to question that article. I did a bit of digging, and the majority seem to say its going to be a new architecture, but nobody seems to be absolutely sure. It certainly will be a smaller process, coming in at 28nm compared to Fermi's 40.

[Asmekiel]

You're right to question that article. I did a bit of digging, and the majority seem to say its going to be a new architecture, but nobody seems to be absolutely sure. It certainly will be a smaller process, coming in at 28nm compared to Fermi's 40.

I guess only time will learn. I'm patient, just got myself a new PC anyway.

[Skelly]

I would imagine its a new architecture.. Nvidia is touting "4x faster than the 500 series". I doubt that could be achieved with a die shrink alone.

[FlawlessSoul]

I would imagine its a new architecture.. Nvidia is touting "4x faster than the 500 series". I doubt that could be achieved with a die shrink alone.

It's 3-4 times faster per watt, which is mostly a result of the die-shrink I believe. If they keep the power usage for their new cards in line with the 500 series then yes, it'll be quite a leap forward.

[haxartus]

If it was a new architecture, we would have heard it by now from Nvidia with presentations and everything. And a die shrink plus a new architecture always creates problems. The last thing they need is a 3-4 months delay with bugged drivers and unoptimized, power hungry cards.
They might call it a new architecture, but it will probably be an improvement on Fermi and that's it