Monday, March 12th 2012

GK104 Silicon Roughly As Big As G92b

When the first PCB shot of the GK104 reference board surfaced, it sent the punters estimating the die area of the GK104 GPU, which the pinned at somewhere around 320 mm². A newer close-up picture of the GK104, helped calculate the figure more accurately, down to around 300 mm² (±5 mm²). This calculation also takes into account that the GK104 chip-package is as big as that of the G92, and the die just as big. It was compared alongside a 55 nm G92b chip. GK104 is NVIDIA's newest performance GPU built on the Kepler architecture, on which SKUs such as the GeForce GTX 680, are based.

Source: Expreview
Add your own comment

34 Comments on GK104 Silicon Roughly As Big As G92b

#1
the54thvoid
cadaveca said:
Dude, we had a thread last week wit hW1zz saying he just went to the Kepler event. LATE LAST WEEK.

If W1zz could travel half-way across the world to get home, build up a rig for the card, get it tested, and write a proper review in like 3 days....man, he'd have to not sleep since starting!


But that laso hleps put things into perspective...W1zz says nVidia gave him a chocolate cookie, and posted a pic. The NDA must expire relatively soon, but I cannot imagine that nV wouldn't give reviewers a week, at least, to get a review done. That puts things at some time next week, iMHO.

Of course, Z77 and IB is supposed to launch soon too, or something...and i haven't the foggiest idea when that is athough I'm supposed to be TPU's board reviewer. Clearly I'm not doing any launch reviews there, nor do I have access right now to unreleased hardware, so perhaps my judgement on NDAs and the time frame given to deal with publishing a reivew is not accurate. Oh well.

GTX570 for $250 local, on that day? That'd be awesome.

G92B stuck around for quite some time. I hope this GPU does as well.
Dave, I respect you but I need to pull you up on a few points:

1) How dare you suggest W1z can't have a review done in that time frame. W1zz exists in abstract time, he's like the lawnmower man. W1zz could swallow stars and make Galactus look skinny.
2) I'm really really impatient.
3) I have money set aside and a few extra water cooling bits and bobs lying homeless until I finish my loop with a shiny new VGA.
4) I've been holding off buying a 7970 for 2 months, waiting to see what Kepler brings. I just needs to knows!!!!
5) I'm really really impatient.

;)
Posted on Reply
#2
cadaveca
My name is Dave
Imsochobo said:
nvidia must have done A.\ something big. B.\ overclocked it alot like amd did not do.
Whne Fermi launched, nV CEO admitted that they used automated tools to make the GPU, and that led to problems with the "interconnecting mesh", or as I understood it, the GPU's cache design. This led to the high power consumption, as cache, in any chip, seems to be the big power draw.

I really think they've taken that Fermi design, and now hand-tooled it, so I am very eager to see how it all works. A fast enough chip, to me, doesn't need as much cache, as long as it's efficient, and give nthe time between the Fermi launch and now should have been more than enough for them to pull out a killer chip for this gen....so much so that I think I'm going nV this time.

I wasn't very interested in the 7970, to be honest. But these chips...and all this power-control stuff, has me very very interested.


the54thvoid said:
Dave, I respect you but I need to pull you up on a few points:

1) How dare you suggest W1z can't have a review done in that time frame. W1zz exists in abstract time, he's like the lawnmower man. W1zz could swallow stars and make Galactus look skinny.
2) I'm really really impatient.
3) I have money set aside and a few extra water cooling bits and bobs lying homeless until I finish my loop with a shiny new VGA.
4) I've been holding off buying a 7970 for 2 months, waiting to see what Kepler brings. I just needs to knows!!!!
5) I'm really really impatient.

;)
I'm just as impatient..I'm just broke. :laugh:


:rockout:
Posted on Reply
#3
Benetanegia
N3M3515 said:
I told you it wouldn't be much faster than 7970 (in the case it is, which i still have my doubts.) but you always over-optimistic with nvidia products.
Yes you told me, but we were discussing about a 340-360 mm^2 chip and not a chip maybe smaller than 300mm^2. It's a very different thing considering it's 256 bit MC in both cases. I was not overoptimistic in any sense. 20% smaller chip == 20% slower chip. So if it's 5% faster than HD7970 with 300mm^2, it would have been 30% faster than HD7970 (so 50% faster than GTX580) if it was as big as leaks suggested when we discussed it. That's MORE than I said, not less.

EDIT: In other words, I always went by transistor efficiency. It should be the same or better than Fermi because of various reasons, and never much worse (because you don't change what is not broken, Fermi was on par with Cayman, better than Tahiti). So when it was at least 340mm^2 that meant 3.6-4.0 billion transistors, which against GF110's 3 billions suggested at least 25-40% more performance. With <300mm^2 it's down to ~3 billion, so expecting it to be much much faster than GF110 starts being more unrealistic, or quite a feat if they manage to do it.
Posted on Reply
#4
N3M3515
Benetanegia said:
20% smaller chip == 20% slower chip. So if it's 5% faster than HD7970 with 300mm^2, it would have been 30% faster than HD7970 (so 50% faster than GTX580)
I agree with that.
Nevertheless, somehow i knew they weren't, going to release a "midrange" part, which could blow 7970 out of the water, regardless of size.
Posted on Reply
#5
phanbuey
N3M3515 said:
I agree with that.
Nevertheless, somehow i knew they weren't, going to release a "midrange" part, which could blow 7970 out of the water, regardless of size.
Well we dont know that yet... and given the driver team of nvidia, they can make magic happen with decent hardware. I am anxious for the reviews.
Posted on Reply
#6
N3M3515
phanbuey said:
Well we dont know that yet... and given the driver team of nvidia, they can make magic happen with decent hardware. I am anxious for the reviews.
One of the first things i learned from hardware launches, never believe in overhyped products.
Posted on Reply