[Entropy]

Originally Posted by Drazick

It's Egg Chicken Circle.
Once they give the performance and capabilities there'll be programs -> applications who would take advantage of that.

I'm not really sure(Don't have any real knowledge in that), but does Video Encoders use DP or do they us SP?

I don't know, for me as a user who doesn't play any game I'm eager to see some applications who would benefit from the GPU, Something beyond gaming.
In much larger scale.

I've been away from the keyboard for a few days, so apologies for the belated response.
Short response is - you'll never see it, if by "In much larger scale" you mean that it will become relevant to the general public.

The reason being that when viewed as a general co-processor the GPU is horribly, horribly inefficient use of transistors, (and thus money and power). At the introduction of Nehalem Intel made quite a bit of propagandistic noise about how new power drawing features had to pay for themselves 2:1 in terms of overall performance or they wouldn't be implemented. And this makes perfect sense as the majority of x86 processors are used in environments where power limits apply, so performance/Watt is critical. Viewed in that light, a GPU is incredibly inefficient even if by magic all x86 software was rewritten from scratch due to the very limited set of problems it can be applied to. In the real world, the level of utilization would be so low so as to be hard to calculate.

So why is it so popular a topic on these boards? Well, GPGPU was something that was pushed by ATI and nVidia in order to try to strengthen their market legs outside gaming, or at least appear to investors as if they did. (I take the somewhat cynical view that this was done to increase the likelyhood of being bought out under good conditions to their shareholders.) So it received a lot of PR attention. It was new, and therefore interesting to those who take an interest in these things.
But it was only ever a valid proposition on the condition that the GPU was already in the system "for free", and any extra use you could put it to was gravy. And this is only true for the core gamer market. For all other users it's essentially true that once you can drive the interface, little more 3D performance is needed. This is why Intel integrated graphics will always suck in the view of the denizens of this forum, and why it should do so. The way the market is moving, the ever increasing focus on power efficiency and cost of the whole system makes it unlikely that a large part of the market will ever have high-power GPUs.

[caveman-jim]

GPGPU is real and has been demonstrated time and again to offer a proven improvement in the computing experience for the end-user. GPU as co-processors is not a pancea for all that ails x86, but it is a great tool for software developers to use to improve their products features and performance. With Direct2D browsers, GPU accelerated Office 2010, GPU accelerated media transcode (Windows 7 drag'n'drop for compatible media foundation devices), and DXVA enabled media players there is already a suite of applications in place.

OpenCL and DirectCompute make it easier for developers to offload their parallel processing workloads from the CPU to the GPU. Like anything there is a lag time from hardware capabilities to mass adoption and mainstream use. Same as FP co-processors, and multiple threads.

Focusing on high-power GPU's is the wrong market - look at the benchmarks for mainstream graphics cards, accelerating performance in say Mediashow. A $100 card can improve the performance of a PC with a $100 processor in it immensely. That's the goal. Not beating a $1300 processor with a $500 card.

[rpg.314]

Originally Posted by Entropy

I've been away from the keyboard for a few days, so apologies for the belated response.
Short response is - you'll never see it, if by "In much larger scale" you mean that it will become relevant to the general public.

The reason being that when viewed as a general co-processor the GPU is horribly, horribly inefficient use of transistors, (and thus money and power). At the introduction of Nehalem Intel made quite a bit of propagandistic noise about how new power drawing features had to pay for themselves 2:1 in terms of overall performance or they wouldn't be implemented. And this makes perfect sense as the majority of x86 processors are used in environments where power limits apply, so performance/Watt is critical. Viewed in that light, a GPU is incredibly inefficient even if by magic all x86 software was rewritten from scratch due to the very limited set of problems it can be applied to. In the real world, the level of utilization would be so low so as to be hard to calculate.

So why is it so popular a topic on these boards? Well, GPGPU was something that was pushed by ATI and nVidia in order to try to strengthen their market legs outside gaming, or at least appear to investors as if they did. (I take the somewhat cynical view that this was done to increase the likelyhood of being bought out under good conditions to their shareholders.) So it received a lot of PR attention. It was new, and therefore interesting to those who take an interest in these things.
But it was only ever a valid proposition on the condition that the GPU was already in the system "for free", and any extra use you could put it to was gravy. And this is only true for the core gamer market. For all other users it's essentially true that once you can drive the interface, little more 3D performance is needed. This is why Intel integrated graphics will always suck in the view of the denizens of this forum, and why it should do so. The way the market is moving, the ever increasing focus on power efficiency and cost of the whole system makes it unlikely that a large part of the market will ever have high-power GPUs.

May be you should look here

http://www.realworldtech.com/page.cf...0909050230&p=2

Your views on perf/W of GPU's might change a bit.

[Jaaanosik]

Originally Posted by CarstenS

How is that? Did I miss an educated die-size estimate or even die foto?

We can now tell you about GF104 die, it is 24.9mm x 14.7mm, and the die size is 367mm^2.

[no-X]

That's bigger than I expected. Tranzistor density is significantly lower compared even to GF100...

[aaronspink]

Originally Posted by no-X

It is expected to be under 400mm² and perform ~20% better than GTX480? It would be nice, but isn't this scenario overly optimistic?

Well, 5870 is smaller and performs better than GF104.

[Alexko]

Originally Posted by no-X

That's bigger than I expected. Tranzistor density is significantly lower compared even to GF100...

It might have something to do with doubled up vias… or perhaps GF104's apparent ability to clock quite high—though obviously not high enough.

[rpg.314]

Well in about 3 years the economics of the GPU market have completely reversed.From the G92 era to the GF104 era.

[mczak]

Originally Posted by no-X

That's bigger than I expected. Tranzistor density is significantly lower compared even to GF100...

It's not too different to GF100. Maybe the difference there is just because there's comparatively more logic & less cache?
GF104 compared to GF100 has:
2/3 of trans
3/4 normal alus
1/1 SFUs / TMUs
2/3 L2 cache (and rops, mem i/o)
1/2 L1 cache (and register file)
1/2 PolyMorph Engines / Raster Engines
Hard to tell though if that's overall really less simple stuff (like cache) percentage-wise...
In any case it's not surprising transistor density is lower than rv870 - that was already very easily seen with g92b / rv770.

Originally Posted by aaronspink

Well, 5870 is smaller and performs better than GF104.

I would argue not by that much though (for a full chip with "reasonable" achievable clocks - 5870 doesn't really overclock too much without additional voltage which makes power draw much worse). Of course, if nvidia could produce a full chip in quantity is another question (I have no idea really).
So if GF104 is 10% bigger and 10% slower basically, that wouldn't be too bad. I would say definitely better than any gt2xx comparisons against rv7xx. Though I suspect GF106 against Juniper will look worse again for nvidia (I still think Cypress didn't scale too well compared to Juniper).

[OlegSH]

Originally Posted by Jaaanosik

We can now tell you about GF104 die, it is 24.9mm x 14.7mm, and the die size is 367mm^2.

Same kind as Neliz's "missing cache structure and etc."(с), except this one from Charly with love

[Alexko]

Originally Posted by mczak

I would argue not by that much though (for a full chip with "reasonable" achievable clocks - 5870 doesn't really overclock too much without additional voltage which makes power draw much worse). Of course, if nvidia could produce a full chip in quantity is another question (I have no idea really).
So if GF104 is 10% bigger and 10% slower basically, that wouldn't be too bad. I would say definitely better than any gt2xx comparisons against rv7xx. Though I suspect GF106 against Juniper will look worse again for nvidia (I still think Cypress didn't scale too well compared to Juniper).

On the other hand and if memory serves, GT200b was a bit better than RV770 in perf/W, while GF104 is a bit worse than Cypress, as the GTX 460 1GB draws 10 to 20W more than the HD 5850 while being a good bit slower.

I wonder how things would look with a full GF104 at 1500MHz or so, compared to the HD 5870.

[chavvdarrr]

I wonder why noone was able to open the chip and take a shot of the die.
Charlie gives some numbers... but no shots :S

[rpg.314]

Originally Posted by chavvdarrr

I wonder why noone was able to open the chip and take a shot of the die.
Charlie gives some numbers... but no shots :S

I don't think getting a die shot is as simple as that. You'll need pretty costly stuff for a die shot, something which will be their with the chip vendors but is unlikely to exist with tech sites/journalists.

[AlexV]

Originally Posted by rpg.314

I don't think getting a die shot is as simple as that. You'll need pretty costly stuff for a die shot, something which will be their with the chip vendors but is unlikely to exist with tech sites/journalists.

To measure die size you need a ruler, and that's exactly what's needed to put this 460 die size thing to rest (that and a bit of time to de-cap it).

[rpg.314]

Originally Posted by AlexV

To measure die size you need a ruler, and that's exactly what's needed to put this 460 die size thing to rest (that and a bit of time to de-cap it).

Yes, But he wasn't angling for the dize size, was he? He seemed a tad more interested int he _die shot_

[AlStrong]

Problem is that you'd only get the top layers of the chip, which may not be particularly interesting. Not sure what sort of lens you'd need. A microscope camera would be nice?

[CarstenS]

Originally Posted by Jaaanosik

We can now tell you about GF104 die, it is 24.9mm x 14.7mm, and the die size is 367mm^2.

Nice rant once again from Charlie. I am looking forward to see which parts turn out to be true in a few months.

While the die size number seems to be a little on the high side of things, I outright doubt that the chip manufacturing and product pricing economics work like Charlie depicts. I don't think, no, I am pretty sure that I couldn't go to TSMCs and buy a wafer for 5000 Dollars, even if that was the price they're asking from Nv or AMD.

edit: Conveniently, he of course forgets that GF104 has also turned of one out of four quad-ROPs/memory controllers/128 kiB L2 Caches, while in Cypress for the higher cost parts, all of those must be present.

Originally Posted by rpg.314

May be you should look here

http://www.realworldtech.com/page.cf...0909050230&p=2

Your views on perf/W of GPU's might change a bit.

I've been adding the data to Davids graph (hope he doesn't mind) for GF100 and Cypress in their respective supercomputing solutions, based on wikipedia data (AMD: 525 DP-GFLOPS, 225 watts, Nvidia: 515 GFLOPS, 225 watts):

http://img682.imageshack.us/img682/1...efficiency.png
This way, it's getting even clearer.

[larrabee]

Originally Posted by rpg.314

I don't think getting a die shot is as simple as that. You'll need pretty costly stuff for a die shot, something which will be their with the chip vendors but is unlikely to exist with tech sites/journalists.

you can use 1500 grit sandpaper and get ok results. it wont be as pretty as the dice you see on the internet but it will work. it's not really a simple project and it would take too much time for a review. they could pay a company like chipworks to do it for them. it takes a lot of skill and background knowledge to do properly.

here's a gallery of DIY deprocessed dies, not mine but pretty cool.
http://ctho.org/gallery/bin/category...ter%20/%20Tech

Originally Posted by AlStrong

Problem is that you'd only get the top layers of the chip, which may not be particularly interesting. Not sure what sort of lens you'd need. A microscope camera would be nice?

if you want the pretty stuff sandpaper and polishing will do the trick.

[Psycho]

Originally Posted by CarstenS

While the die size number seems to be a little on the high side of things, ... Conveniently, he of course forgets that GF104 has also turned of one out of four quad-ROPs/memory controllers/128 kiB L2 Caches, while in Cypress for the higher cost parts, all of those must be present.[/i]

The die size is most probably right, that's easy to get and verify (and if too high nvidia would probably ask some other sites to leak the right number).
He's comparing to 5850/70 combined, so the yield improving by fusing off parts is taking into account, however it's still quite unclear why 5850+70 should yield at 75% and the 460 only at 60%. With the lower density it should have chance of slightly better yield.

Originally Posted by CarstenS

I've been adding the data to Davids graph (hope he doesn't mind) for GF100 and Cypress in their respective supercomputing solutions, based on wikipedia data

Most other numbers are for graphics cards, so the 5870/1G (544@188w) would be more comparable than the 4gb firestream. On the other hand the C2070 is (now) specified at 247w, not 225 (unfortunately no gf100 graphic cards to use here).

[neliz]

Originally Posted by CarstenS

edit: Conveniently, he of course forgets that GF104 has also turned of one out of four quad-ROPs/memory controllers/128 kiB L2 Caches, while in Cypress for the higher cost parts, all of those must be present.

He was always the one hinting that Cypress had a bit more than we would probably ever see.

I've been adding the data to Davids graph (hope he doesn't mind) for GF100 and Cypress in their respective supercomputing solutions, based on wikipedia data (AMD: 525 DP-GFLOPS, 225 watts, Nvidia: 515 GFLOPS, 225 watts):

you used Wikipedia as "facts" on GF100's power usage? 225NVWatts is 325Watt for us.That makes it more close to RV770 in the perf/W figure (seeing as you did pick the right 529mm.

[dkanter]

Originally Posted by neliz

He was always the one hinting that Cypress had a bit more than we would probably ever see.



you used Wikipedia as "facts" on GF100's power usage? 225NVWatts is 325Watt for us.That makes it more close to RV770 in the perf/W figure (seeing as you did pick the right 529mm.

Carsten: I'm actually updating the chart myself, and I hope to put out a revised version on a regular basis. There are a ton of data points that need to be added from both CPU and GPU land.

And I strongly recommend not using wikipedia as a source for anything.

David

[Novum]

Originally Posted by aaronspink

Well, 5870 is smaller and performs better than GF104.

We haven't seen the full version of GF104 yet. There could be yield problems, but I suspect they are holding it back because of strategic reasons.

[eastmen]

Originally Posted by Novum

We haven't seen the full version of GF104 yet. There could be yield problems, but I suspect they are holding it back because of strategic reasons.

The same can be said of cypress

[RecessionCone]

Originally Posted by eastmen

The same can be said of cypress

In what sense? I thought the HD 5870 had all its bits and pieces enabled, and reports seem to indicate that GTX460 has quite a bit more clockspeed headroom. What would a full Cypress look like?

[3dcgi]

Originally Posted by Novum

We haven't seen the full version of GF104 yet. There could be yield problems, but I suspect they are holding it back because of strategic reasons.

What might those strategic reasons be? Low volume due to low yields isn't a strategic reason and I can't think of another reason at the moment.