G80 won't be unified shader architecture

[fierydemise]

Link

Not sure how much impact this'll have and its still early but its news.

 

[jiffylube1024]

Sounds realistic. Also means it will be quicker to the market. 48 pixel shaders eh, why does that sound familiar...?

 

[Cooler]

So G80 not fully direct x10 compatible because unified is a requirement. This reminds me the FX gen problem Microsoft and Nivida did not get it straight on what will be supported in direct x 9.0 and it cost Nivida greatly. If there is one thing they should learn if micosoft say directx version xxxx needss xxxx then you better do or you may suffer.

 

[fliguy84]

who cares as long as it kicks azz

 

[SickBeast]

Originally posted by: fliguy84
who cares as long as it kicks azz

I care if it can't play the next Unreal game or the new DX10 Far Cry game!

 

[rcllbrg]

>or the new DX10 Far Cry game!

My head asplode!

Just thinking about it

 

[tuteja1986]

lol :! go with ATI then

R6XX will be in be awesome and ATI has already developed a unified shader architecture GPU for the Microsoft's XBOX 360. ATI have the head start i guess .

 

[Munky]

Whoah, back up a sec! Who ever said DX10 required unified shaders? Unified shaders is a hardware feature, not a software or an API feature, and it could still be DX10 compliant as long as it supports the required API features even without unified shaders. That being said, it does bring up a question of how effectively it will support the features without unified shaders, and the worst case scenario is it will support it just like the FX series supported DX9. But, it does fit one more piece in the puzzle, and further hints that the g71 will not be a monster card, but rather a transition card, and that NV will try to get the g80 out earlier than Ati can bring out the r600.

 

[otispunkmeyer]

Originally posted by: Cooler
So G80 not fully direct x10 compatible because unified is a requirement. This reminds me the FX gen problem Microsoft and Nivida did not get it straight on what will be supported in direct x 9.0 and it cost Nivida greatly. If there is one thing they should learn if micosoft say directx version xxxx needss xxxx then you better do or you may suffer.

hopefully that wont happen again lol! i remember they packed more into the FX chips than was actually needed. but i remember the R300 was using perfectly acceptable FP24 while the FX could only do FP16 or FP32.....and well it absolutely sucked at FP32

 

[Matt2]

Hopefully this means that G80 will be released before DX10 and Vista comes out. Maybe it'll be Nvidia's last hoo-rah for DX9 and the real monster card for DX10 will be a gen afterwards!

</dream>

 

[nts]

Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

 

[Drayvn]

Originally posted by: munky
Whoah, back up a sec! Who ever said DX10 required unified shaders? Unified shaders is a hardware feature, not a software or an API feature, and it could still be DX10 compliant as long as it supports the required API features even without unified shaders. That being said, it does bring up a question of how effectively it will support the features without unified shaders, and the worst case scenario is it will support it just like the FX series supported DX9. But, it does fit one more piece in the puzzle, and further hints that the g71 will not be a monster card, but rather a transition card, and that NV will try to get the g80 out earlier than Ati can bring out the r600.

IF nVidia are trying to get the leg up on releasing the G80 earlier than the R600 it must be quite a bit. Because ATi have already got a unified shader architecture in place with the Xbox 360 i wouldnt have thought it wouldnt take them very long after to release the R600 early as well.

 

[BenSkywalker]

Why does anyone think that unified shader hardware is better by default?

In absolute terms it is slower, transistor for transistor, without a doubt. The issue comes with load balancing and how much of an impact that is going to have. Unless One architecture is hitting a limitation with one type of shader op and the other type has plenty of free cycles then having a unified model is going to hurt your end performance numbers.

In realistic terms, if the IHV picks a decent combination of pixel/vertex shader units then it will be faster then a unified design- and none of the IHVs will say differently. ATi's focus has been allowing developers flexibility- not that we have seen anything that indicates there is going to be some large shift from the roughly 5:1 balance we have been seeing in terms of pixel to vertex shader load. The only time you are going to see any real benefit is if a dev strays far from the norm in terms of the demands of their engine- we have yet to see this including 360 titles.

Who ever said DX10 required unified shaders? Unified shaders is a hardware feature, not a software or an API feature

Actually, DX10 utilizes unified shaders- it doesn't differentiate between a vertex op and fragment op any longer. The driver can direct calls to the appropriate hardware units(like it does with so many other things) but in terms of coding they are no longer two seperate pieces.

Not sure why this is making news now, nVidia has been saying this for a long time now- NV50(G80) is going to be segmented shader hardware. Last I had heard it was questionable if NV60(likely G100) would move to unified shader hardware.

 

[RussianSensation]

Originally posted by: nts
Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

Although at first 48 pixel pipes seems low, you have to remember these are real pipelines not pixel shaders as in R580. 48 will put it at 2x the pipelines of 7800GTX which is a substantial improvement (since when 2x the performance for new generation is now low?) Also, it's difficult to say how comparable in power a unified shader is to a dedicated shader in your R600 example.

What is interesting is which company is going to catch up first: 1) Will ATI will revamp its OpenGL drivers taking away Nvidia's lustre in those types of games? or 2) Will Nvidia produce a card with more efficient shaders and AA algorithms and take away ATI's shader crown in games like FEAR?

 

[yacoub]

Or will Quake Wars: Enemy Territory get delayed and become a DX10 title like Halo 2 and Unreal Tournament 2007 and whatever else, in an attempt to make the most use of graphical technology and to aid Microsoft's attempt to force gamers onto Vista with all its nasty content lockdowns and neutered abilities short of the Ultimate edition?

If all the good games get delayed, everyone might as well just start saving up for the new DX10 cards coming this Fall and make due with whatever you currently have since anything more would be a waste of money if it already plays FEAR acceptably well since that would then be the most taxing DX9 game released (if, hypothetically, QW:ET were to be delayed).

 

[Steelski]

Originally posted by: RussianSensation

Originally posted by: nts
Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

Although at first 48 pixel pipes seems low, you have to remember these are real pipelines not pixel shaders as in R580. 48 will put it at 2x the pipelines of 7800GTX which is a substantial improvement (since when 2x the performance for new generation is now low?) Also, it's difficult to say how comparable in power a unified shader is to a dedicated shader in your R600 example.

What is interesting is which company is going to catch up first: 1) Will ATI will revamp its OpenGL drivers taking away Nvidia's lustre in those types of games? or 2) Will Nvidia produce a card with more efficient shaders and AA algorithms and take away ATI's shader crown in games like FEAR?

I dont really think you know what you are propositioning with saying that the next chip will have 48 normal pipelines. You are talking about a chip that is much much larger. and the fact that 0.65 processes do not exsist in the GPU world yet makes this a very very very expensive card by default. even with a 0.80 process this is still so huge its not even worth thinking about. most likley they will have 24 real pipes and 2 alu's per pipe. what currant games indicate is that there is no need for more Texture units than the currant 24 so i dont think Nvidia will be silly enough to implement 48 of them at great cost.
What this news really indicates is that ATI are very much in the driving seat next round and its theirs to looooz.
What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.
Another thing to take on board is that it pays to be good to Microsoft.

 

[BenSkywalker]

What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.

Speaking on a computational level this sounds, less then fully informed to be polite as possible. It would be akin to saying that a processor is two and a half generations behind because it has seperate SIMD and FPU units while another has them combined.

Or put another way, you could say that the R480 was a generation behind the original GeForce since it had angle dependant AF while the GF didn't

 

[Genx87]

Originally posted by: Cooler
So G80 not fully direct x10 compatible because unified is a requirement. This reminds me the FX gen problem Microsoft and Nivida did not get it straight on what will be supported in direct x 9.0 and it cost Nivida greatly. If there is one thing they should learn if micosoft say directx version xxxx needss xxxx then you better do or you may suffer.

afaik having a Unified Shader pipeline is not a requirement of DX10.

 

[jim1976]

G80 = geometry + vertex shaders unified, PS separate, "G90" complete USC, no ROPs AFAIK
And as Munky and some others correctly stated USC is not a "must" in order to be d3d10 compliant...
As far as I know R600 might have more than 64ALUS with 4madd max each..
The bet still remains to see if R600's monsterous theoretical numbers will have the practical results that ATI aims. Will USC be sufficient for high performance in the forthcoming games?

 

[Steelski]

Originally posted by: BenSkywalker

What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.

Speaking on a computational level this sounds, less then fully informed to be polite as possible. It would be akin to saying that a processor is two and a half generations behind because it has seperate SIMD and FPU units while another has them combined.

Or put another way, you could say that the R480 was a generation behind the original GeForce since it had angle dependant AF while the GF didn't

Nevermind. Deny what I say all you want. The fact that it will be about that behind in experience with Unified shaders is not a small thing but i know its not the same as saying it will be 2.5 gens behind in performance. but considering that ATI is likley to have provided many unified chips before Nvidia shows one makes their unified shader program with little experience whilst ATI would have plenty to make an efficient chip. Can you deny that? you can compare it to ati and SM3.0. they were a gen behind last generation. Unfortunatley Nvidia dident really add anything to this gen and allowed ATI to catch up quick enough. this could happen with the unified shaders aswell.

 

[SpeedZealot369]

Originally posted by: yacoub
Or will Quake Wars: Enemy Territory get delayed and become a DX10 title like Halo 2 and Unreal Tournament 2007 and whatever else, in an attempt to make the most use of graphical technology and to aid Microsoft's attempt to force gamers onto Vista with all its nasty content lockdowns and neutered abilities short of the Ultimate edition?

If all the good games get delayed, everyone might as well just start saving up for the new DX10 cards coming this Fall and make due with whatever you currently have since anything more would be a waste of money if it already plays FEAR acceptably well since that would then be the most taxing DX9 game released (if, hypothetically, QW:ET were to be delayed).

 

[Acanthus]

Originally posted by: Steelski

Originally posted by: RussianSensation

Originally posted by: nts
Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

Although at first 48 pixel pipes seems low, you have to remember these are real pipelines not pixel shaders as in R580. 48 will put it at 2x the pipelines of 7800GTX which is a substantial improvement (since when 2x the performance for new generation is now low?) Also, it's difficult to say how comparable in power a unified shader is to a dedicated shader in your R600 example.

What is interesting is which company is going to catch up first: 1) Will ATI will revamp its OpenGL drivers taking away Nvidia's lustre in those types of games? or 2) Will Nvidia produce a card with more efficient shaders and AA algorithms and take away ATI's shader crown in games like FEAR?

I dont really think you know what you are propositioning with saying that the next chip will have 48 normal pipelines. You are talking about a chip that is much much larger. and the fact that 0.65 processes do not exsist in the GPU world yet makes this a very very very expensive card by default. even with a 0.80 process this is still so huge its not even worth thinking about. most likley they will have 24 real pipes and 2 alu's per pipe. what currant games indicate is that there is no need for more Texture units than the currant 24 so i dont think Nvidia will be silly enough to implement 48 of them at great cost.
What this news really indicates is that ATI are very much in the driving seat next round and its theirs to looooz.
What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.
Another thing to take on board is that it pays to be good to Microsoft.

So with this massive paragraph you posted. Do you honestly believe TSMC and UMC wont be at 65nm in 6 months? come on now. (which would allow double the transistors in the same die size, at a higher clockspeed)

 

[jim1976]

Originally posted by: Acanthus

Originally posted by: Steelski

Originally posted by: RussianSensation

Originally posted by: nts
Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

Although at first 48 pixel pipes seems low, you have to remember these are real pipelines not pixel shaders as in R580. 48 will put it at 2x the pipelines of 7800GTX which is a substantial improvement (since when 2x the performance for new generation is now low?) Also, it's difficult to say how comparable in power a unified shader is to a dedicated shader in your R600 example.

What is interesting is which company is going to catch up first: 1) Will ATI will revamp its OpenGL drivers taking away Nvidia's lustre in those types of games? or 2) Will Nvidia produce a card with more efficient shaders and AA algorithms and take away ATI's shader crown in games like FEAR?

I dont really think you know what you are propositioning with saying that the next chip will have 48 normal pipelines. You are talking about a chip that is much much larger. and the fact that 0.65 processes do not exsist in the GPU world yet makes this a very very very expensive card by default. even with a 0.80 process this is still so huge its not even worth thinking about. most likley they will have 24 real pipes and 2 alu's per pipe. what currant games indicate is that there is no need for more Texture units than the currant 24 so i dont think Nvidia will be silly enough to implement 48 of them at great cost.
What this news really indicates is that ATI are very much in the driving seat next round and its theirs to looooz.
What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.
Another thing to take on board is that it pays to be good to Microsoft.

So with this massive paragraph you posted. Do you honestly believe TSMC and UMC wont be at 65nm in 6 months? come on now. (which would allow double the transistors in the same die size, at a higher clockspeed)

m8 do not underestimate the difficulty of going to 65nm... Mark my words.. I'm not saying that it is not possible I'm saying it is difficult

 

[Steelski]

yes i do! they are not even using 0.80 at the moment so I would be willing to put a wager on it. traditionally CPU's all go to the next step a while before GPU's. I see Intel but not AMD yet. when that happens i would be more than happy to say that they will.
What really swayes my judgement on this though is the rumour that ATI will be using .80 for their next process (due out after the G80), of course this is just rumour. but we shal see.

 

[Acanthus]

Originally posted by: jim1976

Originally posted by: Acanthus

Originally posted by: Steelski

Originally posted by: RussianSensation

Originally posted by: nts
Interesting. If true then they probably took the TMUs out of the Pixel pipes and are sharing them between the vertex and pixel units.

ATi is definately ahead here then, hopefully won't turn into another FX generation but it looks like they are going for insane clocks for this chip (48 pixel pipes seems low).

What's the R600 rumor, 64 unified shaders + 16/24/32 texture units?

Although at first 48 pixel pipes seems low, you have to remember these are real pipelines not pixel shaders as in R580. 48 will put it at 2x the pipelines of 7800GTX which is a substantial improvement (since when 2x the performance for new generation is now low?) Also, it's difficult to say how comparable in power a unified shader is to a dedicated shader in your R600 example.

What is interesting is which company is going to catch up first: 1) Will ATI will revamp its OpenGL drivers taking away Nvidia's lustre in those types of games? or 2) Will Nvidia produce a card with more efficient shaders and AA algorithms and take away ATI's shader crown in games like FEAR?

I dont really think you know what you are propositioning with saying that the next chip will have 48 normal pipelines. You are talking about a chip that is much much larger. and the fact that 0.65 processes do not exsist in the GPU world yet makes this a very very very expensive card by default. even with a 0.80 process this is still so huge its not even worth thinking about. most likley they will have 24 real pipes and 2 alu's per pipe. what currant games indicate is that there is no need for more Texture units than the currant 24 so i dont think Nvidia will be silly enough to implement 48 of them at great cost.
What this news really indicates is that ATI are very much in the driving seat next round and its theirs to looooz.
What could also be seen is that the round after G80 and R600 will most likley be unified shaders.....Where will Nvidia be there. 2.5 generations behind with that technology when you consider that ATI will have the R500,R600 and the R600 refresh.
Another thing to take on board is that it pays to be good to Microsoft.

So with this massive paragraph you posted. Do you honestly believe TSMC and UMC wont be at 65nm in 6 months? come on now. (which would allow double the transistors in the same die size, at a higher clockspeed)

m8 do not underestimate the difficulty of going to 65nm... Mark my words.. I'm not saying that it is not possible I'm saying it is difficult

The research is already done elsewhere all they have to do is retool.