Nvidia reveals Specifications of GT300

[cbn]

http://www.brightsideofnews.co...-revealed---its-a-cgpu!.aspx

I couldn't get the link to work but here is what it said:

What is GT300?

Even though it shares the same first two letters with GT200 architecture [GeForce Tesla], GT300 is the first truly new architecture since SIMD [Single-Instruction Multiple Data] units first appeared in graphical processors.

GT300 architecture groups processing cores in sets of 32 - up from 24 in GT200 architecture. But the difference between the two is that GT300 parts ways with the SIMD architecture that dominate the GPU architecture of today. GT300 Cores rely on MIMD-similar functions [Multiple-Instruction Multiple Data] - all the units work in MPMD mode, executing simple and complex shader and computing operations on-the-go. We're not exactly sure should we continue to use the word "shader processor" or "shader core" as these units are now almost on equal terms as FPUs inside latest AMD and Intel CPUs.

GT300 itself packs 16 groups with 32 cores - yes, we're talking about 512 cores for the high-end part. This number itself raises the computing power of GT300 by more than 2x when compared to the GT200 core. Before the chip tapes-out, there is no way anybody can predict working clocks, but if the clocks remain the same as on GT200, we would have over double the amount of computing power.
If for instance, nVidia gets a 2 GHz clock for the 512 MIMD cores, we are talking about no less than 3TFLOPS with Single-Precision. Dual precision is highly-dependant on how efficient the MIMD-like units will be, but you can count on 6-15x improvement over GT200.


This is not the only change - cluster organization is no longer static. The Scratch Cache is much more granular and allows for larger interactivity between the cores inside the cluster. GPGPU e.g. GPU Computing applications should really benefit from this architectural choice. When it comes to gaming, the question is obviously - how good can GT300 be? Please do bear in mind that this 32-core cluster will be used in next-generation Tegra, Tesla, GeForce and Quadro cards.

This architectural change should result in dramatic increase in Dual-Precision performance, and if GT300 packs enough registers - performance of both Single-Precision and Dual-Precision data might surprise all the players in the industry. Given the timeline when nVidia begun work on GT300, it looks to us like GT200 architecture was a test for real things coming in 2009.

Just like the CPU, GT300 gives direct hardware access [HAL] for CUDA 3.0, DirectX 11, OpenGL 3.1 and OpenCL. You can also do direct programming on the GPU, but we're not exactly sure would development of such a solution that be financially feasible. But the point in question is that now you can do it. It looks like Tim Sweeney's prophecy is slowly, but certainly - coming to life.

 

[akugami]

Your link is screwed up.

nVidia GT300 Link

EDIT:

@Just Learning

Heh, you copy and pasted that before I threw up the link for those who want to check out where it came from. Anyways, what happened is the forum software assumed you were finished with your sentence (the ! mark at the end) and cut off the last bit of your URL, specifically the "!.aspx" at the end of the link.

 

[cbn]

Originally posted by: akugami
Your link is screwed up.

nVidia GT300 Link

Thanks that is the article I am talking about.

 

[OCGuy]

Wow...that could be an amazing chip. :Q

 

[Tempered81]

Performance increase going GT200->GT300 is similar to percentage increase in RV670->RV770 in terms of shader count. But now that the shaders are redesigned, it could actually be double or triple the performance.... 900 mflops to 3 tflops.

 

[SunnyD]

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

 

[cbn]

What is "dual precision" going to be used for?

 

[OCGuy]

Originally posted by: SunnyD

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

Nice attempt to threadcrap by drawing conclusions which you have no data to back-up



3Tflops would be amazing. I cant wait for the next gen...

 

[Idontcare]

Is GT300 intended to be a 40nm part or a 55nm part?

 

[cbn]

Originally posted by: Idontcare
Is GT300 intended to be a 40nm part or a 55nm part?

I am sure it will be 40nm.

P.S. A 360 GTX with 448 stream processors (rather than 512) sounds pretty good to me. If ATI comes out with HD58xx first this may even have the chance at being reasonably priced (at least in a relative sense)

 

[alyarb]

it will definitely be 40nm. this sounds crazy.

 

[taltamir]

this, makes me think of an article i read about heterogeneous computing, and how it requires a distinct departure from both CPU and GPU design... where all cores scale 100% all the time, they don't have to be the same type of core, they just process data instead of threads... I don't think its there yet, but i think it is a step in that direction. IF nvidia continues that way they really COULD break the x86 stranglehold... (and replace it with the nvidia stranglehold)

 

[cbn]

So will ATI 58xx have to come out with 2000 stream processors to compete? Or will they continue to stick with a similar die size?

I am betting 1600 stream processors will be enough as long as ATI comes out with HD58xx before Nvidia releases GT300.

 

[aka1nas]

Originally posted by: Just learning
What is "dual precision" going to be used for?

It's needed on the GPGPU side in particular for a lot of HPC applications that need more than single precision (we're basically talking about how many places after a decimal point you can have). Single precision is insufficient for certain applications that use very precise decimal values and cannot tolerate any errors introduced from rounding or truncation.

On the gaming side, IIRC Carmack's been asking for floating point color values for a while now.

 

[nitromullet]

Originally posted by: OCguy

Originally posted by: SunnyD

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

Nice attempt to threadcrap by drawing conclusions which you have no data to back-up



3Tflops would be amazing. I cant wait for the next gen...

While it's certainly an attempt to threadcrap, you can't honestly say that it probably isn't true. That aside huge, hot, power hungry, and expensive has never stopped me before provided the performance is there.

 

[cbn]

Originally posted by: taltamir
this, makes me think of an article i read about heterogeneous computing, and how it requires a distinct departure from both CPU and GPU design... where all cores scale 100% all the time, they don't have to be the same type of core, they just process data instead of threads... I don't think its there yet, but i think it is a step in that direction. IF nvidia continues that way they really COULD break the x86 stranglehold... (and replace it with the nvidia stranglehold)

At that point are we talking about true integrated GPU/CPU as a single unit? If so wouldn't or couldn't that reduce other total power consumptions related to the need for other accessory chipsets?

Also do you know how this "dual precision" relates things?

 

[Keysplayr]

Originally posted by: SunnyD

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

You don't know the size, you don't know the heat dissapation, you don't know the power it will draw, you don't know the price. Thanks for crapping by.

 

[alyarb]

i'd imagine they are going for a ~250-280 watt TDP. who in their right minds would exceed 300w for one chip?

 

[Idontcare]

Originally posted by: Just learning

Originally posted by: Idontcare
Is GT300 intended to be a 40nm part or a 55nm part?

I am sure it will be 40nm.

P.S. A 360 GTX with 448 stream processors (rather than 512) sounds pretty good to me. If ATI comes out with HD58xx first this may even have the chance at being reasonably priced (at least in a relative sense)

Cool. What is the ETA?

 

[thilanliyan]

Originally posted by: Keysplayr

Originally posted by: SunnyD

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

You don't know the size, you don't know the heat dissapation, you don't know the power it will draw, you don't know the price. Thanks for crapping by.

They're pretty good guesses though. If what he said was completely unfathomable (like saying GT300 would perform worse than GT200) I could see your issue with it. But can you honestly say GT300 (and the ATI 5000 series for that matter) WON't be larger, and more power hungry than this generation's cards?

I personally don't have an issue with it though...low power draw has not been that high my list...as long as I can have low IDLE power consumption I'm all good. It seems to be the price of progress.

 

[alcoholbob]

GTX 295 on a single chip? W00t

 

[Idontcare]

Originally posted by: thilan29

Originally posted by: Keysplayr

Originally posted by: SunnyD

Originally posted by: OCguy
Wow...that could be an amazing chip. :Q

Amazingly HUGE and HOT and POWER HUNGRY... yeah. Oh yeah, also amazingly EXPENSIVE too.

You don't know the size, you don't know the heat dissapation, you don't know the power it will draw, you don't know the price. Thanks for crapping by.

They're pretty good guesses though. If what he said was completely unfathomable (like saying GT300 would perform worse than GT200) I could see your issue with it. But can you honestly say GT300 (and the ATI 5000 series for that matter) WON't be larger, and more power hungry than this generation's cards?

55nm -> 40nm transition involved in there too, which makes most assertions regarding power consumption and die-size a pointless debate until we have data.

 

[HOOfan 1]

I am sure when super sports cars are designed, the fuel consumption and price are at the bottom of the list of priorities.

If you want a cool, cheap power saving chip then don't buy something like this or the HD4870X2 or the GTX 295.

If you want something that is fast for gaming...buy one of those.

Personally, when I select a graphics card performance is my number 1 priority. If it is power hungry, I buy a better PSU. If it is hot, I put it in a case with better airflow. If it is expensive, I eat crackers for lunch instead of fresh sliced deli meat.

 

[thilanliyan]

Originally posted by: Idontcare
55nm -> 40nm transition involved in there too, which makes most assertions regarding power consumption and die-size a pointless debate until we have data.

I wouldn't dream of really debating without the data . I'm curious to see whether 40nm brings any power consumption benefits. I guess we'll find out when ATI's RV770 comes out in a couple of weeks.

I was just stating what we've had historically...which has been larger GPUs and increased power consumption with successive architectures. Someone correct me if I'm wrong but I don't think there's been any new major architecture changes that have come with a smaller die and lower power consumption.

 

[alyarb]

die shrinks always bring power consumption benefits, which is why the GTX 285 consumes 40-50 fewer watts than the GTX 280. the idea with GT300 is that we have more room to pack more functional units, so that for a given TDP we get more flops or whatever your metric is.