NVIDIA GF100 & Friends speculation

Why? In small sequential access case, it's should be close to full bandwidth (larger sequential access case will have problems with timing, since other memory controllers will have to wait for the one with bigger memory). In random access case, the memory controller with big memory does not have to always do two bursts since it's quite possible that one burst can cover the data you want and you don't have to do both bursts.

Of course, we don't know for sure how NVIDIA does it without further testing. It's possible that NVIDIA did it the simple way and put all extra memory in a single memory controller. Since the CUDA bandwidth test doesn't seem to tell much (although, if the test does very large sequential access, it could hit timing problem if the memory is interleved as said in my post), I think we'll have to wait for more data to find out.

[EDIT] I think there are two cases here: the first one is that CUDA only exposes ~768MB memory, which suggests that NVIDIA uses a "flex" model (all extra memory in a single controller), and NVIDIA does not want to expose the slower memory region to GPGPU applications. However, another one is the CUDA bandwidth test, which seems to suggest that the effective bandwidth is the same as a 450 (clock normalized). This seems to suggest a full interleave model.

I think memory management for a "flex" model is probably not very hard. NVIDIA may simply make the extra memory as some kind of "faster Turbo cache." They already have management code for TC so it's easier to reuse them. The downside of a full interleave model is that when you are actually doing a lot of sequential memory access you'll have to wait for the slowest one (i.e. the memory controller with the bigger memory) and it will bring down performance to roughly equivalent to a 128 bits bus. Although in general 3D rendering does not do a lot of large sequential access, sometimes it still needs to do them especially in deferred shaders.

 

Some GTX590 rumors:

Nvidia has signed a pact with the devil, otherwise how else can the performance of GeForce GTX 590 be explained!? On the second day I cannot understand how it is possible that the GeForce GTX 590 is more powerful than two GeForce GTX 580 at the same frequencies (generally speaking, they, should be 500 MHz) ... is this a newer and better driver? or improved efficiency through the on-pcb PCI-E switch? I just do not get it ... but that is how it is

http://obrovsky.blogspot.com/2011/03/geforce-gtx-590-in-details.html

UPDATE: First impressions are great, this is quietest high-end card iv ever seen, quieter then GTX 580! GeForce logo on cooler lights!!! Performance is incredible too! Have we a new king ...?

 

some things in there made me lol (a lot.. and loud!)

 

No wonder, it comes from famous OBR

 

Nvidia GTX 590 specs?

Core Clock 752MHz
Memory Clock 1002MHz
Memory Bus 384bit
Memory Size 3072MB
http://www.techpowerup.com/gpudb/281/.html

 

False, the clocks are 607-613/850-855MHz (depending on sources)
At TPU clocks, they wouldn't get anywhere near 365W TDP, not even with the "Fermi TDP" formula

 

Maybe that's their "Awesome" mode When is the reveal supposed to be?

 

Well.. if "average tdp" includes all the time you're not gaming

 

ASUS GTX 590 Listed
http://www.expc.ca/i-17979-ASUS_ENGTX590_3DIS_3GD5.html

 

and a bargain too

 

Usually factory overclocked cards a bit more expensive

 

Lol the dawn of the percent point factory OC era.

 

WOOOOOO! HELL YEAH!

Finally, Nvidia releases a performance/enthusiast class card with a DP output (assuming those specs are real). About freaking time. All their cards from this point forward had better feature at least one DP output, I hope.

The card itself is completely irrelevant to me, but it's nice to see DP finally appearing on Nvidia cards.

Regards,
SB

 

Actually there's GTX580's with DP too, and probably others too - wether they're part of reference design on 590 either is still unknown

 

MSI GTX 590
http://cdn2.techbang.com.tw/posts/5...the-fastest-graphics-card-msi-n590gtx-p3d3gd5
Translated
http://translate.google.com/transla...the-fastest-graphics-card-msi-n590gtx-p3d3gd5

 

Ah, well that'll suck and be a shame if it's not standard on all cards. Keep hoping Nvidia will step up and stop lagging behind in these things.

Regards,
SB

 

http://www.techpowerup.com/forums/showpost.php?p=2231230&postcount=85

 

Wins on looks and size. Just the important stuff left now.

 

Doesn't look like that 13 MHz extra is going to save it.

 

I guess with these clocks , the HD 6990 lead is inevitable .