How can Nvidia be ahead of ATI but 360 GPU is on par with RSX?

Status
Not open for further replies.
macabre said:
Sorry I can`t remember exactly but wasn`t it mentioned on these boards that cell could at least render certain parts of the frame or apply framebuffer effects ?
Hmmm... Well, I'm sure many people could argue that the reason that Cell is such a FP monster is because Sony wanted it to handle some graphics tasks. I hope these aren't the same folks who think that somehow Cell = PPU as well. You can't have it both ways. Hopefully devs will have freedom on how they use Cell.
 
Alpha_Spartan said:
Hopefully devs will have freedom on how they use Cell.

What a silly statement, if you don't mind me saying. Of course they will.

The "if Cell is used for graphics then it won't have enough power for anything else/will only be as powerful as Xenon" statements being bandied about elsewhere are also ridiculous.
 
CELL for graphics

Alpha_Spartan said:
Hmmm... Well, I'm sure many people could argue that the reason that Cell is such a FP monster is because Sony wanted it to handle some graphics tasks. I hope these aren't the same folks who think that somehow Cell = PPU as well. You can't have it both ways. Hopefully devs will have freedom on how they use Cell.

EA is using 3 SPEs for graphics, 4 for physics, AI, etc. Fill-rate limited. Good situation for developer on closed box.
 
Alpha_Spartan said:
Hmmm... Well, I'm sure many people could argue that the reason that Cell is such a FP monster is because Sony wanted it to handle some graphics tasks. I hope these aren't the same folks who think that somehow Cell = PPU as well. You can't have it both ways. Hopefully devs will have freedom on how they use Cell.

Look Alpha I'm going to say this once. I'm not PM'ing you on this either because I'm not going to have my rep looking bad because you keep trying to put me down in front of everybody. I can only go by what you say. And at first you were up front and direct. You said the RSX will prove that it is a 11th hour G70 put into the PS3. And then you went back and change it a little.

I'm cool with you changing it. You should have just said my bad or I'm sorry I meant to say if. Not blast at me. I thought you were some guy trying to put down the RSX. But it looks like you are not. Just try to explain yourself a little better next time.

PEACE.
 
Uhh, the RSX is an 11th hour move by Sony is it not? Was it planned to be in the PS3 all along? No. Sources clearly say Sony was working on its own internal GPU and maybe one from Toshiba also. That didn't pan out, so Sony ended up looking elsewhere as soon as it could and found Nvidia was offering their tech to put in PS3. Maybe it didn't happen in the matter of a few months, but it wasn't really that long ago.
 
DaveBaumann said:
Errr, this is a discussion about the graphics chips, as the title quite clearly states, not the verall system - this element is very much a factor of the different graphics chips withing these parts.

Oh come on Dave, don't you think I know the thread title and the context of the discussion?! Anyway it was a natural extension on your load balancing theme and was a precursor to my subsequent comments.

It's clear from the architecture that the pool of SPUs are the load balancing resources in PS3, to aid the PPE and RSX as required. Where as Xenos' pool of ALUs act as X360's load balancing resource, especially with MemeExport...

DaveBaumann said:
And this is likewise with Xenos, except the curiousity is that if you move workload such as vertex processing from Xenos to the CPU, Xenos gains resources to spend on pixel shading.

That curiosity was one of my first thoughts when the G70 pixel shaders were disclosed and hence the earlier comparison under those conditions. Suffice to say, CELL and XeCPU will have different resources available under those conditions...

SubD said:
The 'overall system' is the 'graphic chip' for the PS3.

Well, I kinda see it the same way too, i.e. CELL-RSX act like SLI'ed GPUs with different traits. CELL behaving more like a GPU than a CPU...
 
Titanio said:
What a silly statement, if you don't mind me saying. Of course they will.
IIRC, PSX developers didn't have this freedom. It's only with the Playstation 2 that devs were "allowed" (*cough* had no choice *cough*) to see the bare metal. So it's not as clear cut and common sense as you believe.
The "if Cell is used for graphics then it won't have enough power for anything else/will only be as powerful as Xenon" statements being bandied about elsewhere are also ridiculous.
Is this in reference to my PPU comment? I seriously believe that if the Cell is implemented as a graphics helper that we will have another EE situation where the majority of FP processing was devoted to graphics and so in real-world situations the EE fared no better than the Xbox's Celeron in terms of overall performance.
 
Sonic said:
Uhh, the RSX is an 11th hour move by Sony is it not? Was it planned to be in the PS3 all along? No. Sources clearly say Sony was working on its own internal GPU and maybe one from Toshiba also. That didn't pan out, so Sony ended up looking elsewhere as soon as it could and found Nvidia was offering their tech to put in PS3. Maybe it didn't happen in the matter of a few months, but it wasn't really that long ago.

Still doesn't explain why Sony had a relationship with Nvidia with regards to making a GPU since late 2002.
 
mckmas8808 said:
Still doesn't explain why Sony had a relationship with Nvidia with regards to making a GPU since late 2002.

Has that fully been explained? I haven't seen an article discussing the full extent of Sony's and Nvidias relationship or when they started to "hook up" :LOL:. Would you happen to have a link to something like that mckmas? I would like to read about it.
 
Jaws said:
It's clear from the architecture that the pool of SPUs are the load balancing resources in PS3, to aid the PPE and RSX as required.
I haven't seen anything that suggests SPEs load balance for the RSX in the same manner that Xenos' ALUs load balance between themselves. Xenos does so on a much finer grain level than I would ever expect from SPEs. Is there some article I missed on the topic? Unless there is, I really don't think what you're advocating is really that comparable to what Dave is talking about.
 
BlueTsunami said:
Has that fully been explained? I haven't seen an article discussing the full extent of Sony's and Nvidias relationship or when they started to "hook up" :LOL:. Would you happen to have a link to something like that mckmas? I would like to read about it.

Of course I don't have any full 20 page reports on what they have and have not been doing together. I read the samething that 80% of the guys here have read. In December of 2004 they stated that they started to work together two years ago. Just because I don't have a 20 page report doesn't mean that it didn't happen.
 
Alpha_Spartan said:
IIRC, PSX developers didn't have this freedom. It's only with the Playstation 2 that devs were "allowed" (*cough* had no choice *cough*) to see the bare metal. So it's not as clear cut and common sense as you believe.

I'm sorry, but you seriously think Sony is reserving certain resources for certain areas of a game? What?

Alpha_Spartan said:
Is this in reference to my PPU comment? I seriously believe that if the Cell is implemented as a graphics helper that we will have another EE situation where the majority of FP processing was devoted to graphics

Why would it be a majority? Why would one or two SPUs used for graphics, for example, mean the majority of its power was being spent on graphics?

The assumption that SPUs do one thing and never do anything else is also very prevalent and misleading btw.

Alpha_Spartan said:
and so in real-world situations the EE fared no better than the Xbox's Celeron in terms of overall performance.

This, relatively common analogy in some quarters, with regard to Cell "helping" with graphics is so wrong on so many levels. They're not remotely comparable situations.
 
Wouldn't it hurt PS3's general processing power if 2 or 3 SPU's are used for graphics? Why would this be so desirable? Wouldn't it be better if the GPU (which was designed for the task) would do, say 95 percent of the graphics?
 
hadareud said:
Wouldn't it hurt PS3's general processing power if 2 or 3 SPU's are used for graphics? Why would this be so desirable?

Because you could do certain things more easily than any GPU could. Or you could take some load off the GPU to let it do even more on its own.

It would reduce the power for other things, but there is plenty of power there, and it happens to suit graphics work amongst other things.
 
hadareud said:
Wouldn't it hurt PS3's general processing power if 2 or 3 SPU's are used for graphics? Why would this be so desirable? Wouldn't it be better if the GPU (which was designed for the task) would do, say 95 percent of the graphics?

yes it would be more desirable for gpu to do gpu work, isn't that the point of moving to dedicated processors?
 
TTBOMK there's never been any such info other then the offical nVidia spokesperson's 'we're pleased to have been working closely with Sony for 2 years'. OR soundbites are the only info we have on PS3's GPU. Though didn't one's translation of the Goto article include some GPU references?

Sonic : I don't know that not being planned from the start = last minute. AFAIK several options were considered and the best chosen. At least that's an option. Perhaps 2 years ago Sony, Toshiba and nVidia all presented GPU solutions, built a few prototypes as needed, and finally Sony chose their GPU for PS3? Nothings certain in this context.

As for nVidia and Sony's relationship, even if last minute it sounds, at least from nVidia's POV, like a fortuitous relationships. nVidia have said they've worked with other companies, but share a vision with Sony. Could just be hyped marketting speak of course, but what they say, and the announcement of further collaborations, would suggest nVidia is (or set to become) more then just an emergency GPU supplier.
 
dukmahsik said:
yes it would be more desirable for gpu to do gpu work, isn't that the point of moving to dedicated processors?

That was what I was trying to say. It just seemed to me that a few people used the cell as a "defense" for RSX (not that it needs defending, it seems a very capable GPU) ...
 
Shifty Geezer said:
TTBOMK there's never been any such info other then the offical nVidia spokesperson's 'we're pleased to have been working closely with Sony for 2 years'. OR soundbites are the only info we have on PS3's GPU. Though didn't one's translation of the Goto article include some GPU references?.

ya that's all there's ever been, a very vague "we have been working with..." statement.

Alot of people have taken that as proof that the RSX has been in development for 2-3 years.
 
dukmahsik said:
yes it would be more desirable for gpu to do gpu work, isn't that the point of moving to dedicated processors?
It's a very simpl,e equation. You have x amount of processing capacity in the console made of a CPU and a GPU, which has to perform certain tasks like rendering, shading, AI, physics, game logic. You divvy these tasks between the resources you have to maximise them. If overall using half the CPU for graphics helps the overall system, you do it. You don't grumble 'oh no, we're having to use the CPU to do graphics work.' Likewise if you can use the GPU to aid physics calculations, you don't refuse to use it on the principle it's a GPU and shouldn't be used for anything other than graphics.

If a developer wants to use some of Cell's power to help get better graphics, why shouldn't they? And it eats into Cell's FLOPs does it? Goodness, flops are there to be used! It's not like there's a shortage of processing power and a few million vertex tranforms is going to take 95% of CPU time!
 
hadareud said:
That was what I was trying to say. It just seemed to me that a few people used the cell as a "defense" for RSX (not that it needs defending, it seems a very capable GPU) ...

It's not a defense, it just happens that it can cover some concerns raised re. RSX - and then a whole lot more too. When you have a CPU sitting at the end of a very fat pipe that happens to be very good at "graphics", it's difficult to ignore. Its potential is a lot greater than meeting some checklist.

This harks back to the debate over context, and discussing these things as systems. Personally I don't think you can meaningfully discuss any one component in complete isolation in these systems. They do tend to stand tall on their own, thank goodness, but that's useless without context.
 
Status
Not open for further replies.
Back
Top