A PS3 breakdown of all components.

Status
Not open for further replies.
I have done significantly more research on the Xenos and the unified shader architecture is still hazy, but I understand it a bit better. I also am very interested in exactly how exploiting the MEMEXPORT feature of the Xenos works out to aid the rest of the system performance of the XBox360. I am slowly starting to get a better understanding of how things flow inside of a graphics card alone so I can make a better assessment or judgement to compare the cards. Although, if I get to a much higher level of understanding the hardware before any more information on the RSX comes out, it's likely I'll determine it impossible to non-speculatively judge the performance of both consoles before the PS3 comes out and software is actually written for it altogether.

@Acert93 - perhaps you missed the section of my post where I explicity say that better hardware doesn't make better games. Perhaps you also missed my prior post which also clearly shows that the stance is for software to evolve. It comes down to the games and what the developers actually do, but honestly, you can't judge the value of any processor by the worst applications that run on it. So it makes more sense to talk about the potential.

Stop trying to call out my bias please. Bias is admittedly there so let's aim on making it better, not point out where it's wrong or biased. That is what leads to a flame war. Talk about the merits, not the bias. All people have bias no matter what they are talking about. If the bias supports false(or weakly true) reasoning and logic, you can disqualify it using correct(or stronger) logic or reasoning alone. If the bias supports correct(on the deepest levels of) logic and reasoning, then all you're getting mad at is the bias and the only thing left to do is flame the bias holder even if he is right. Or, people end up trying to argue against correct logic, with incorrect logic...is the cycle making sense?

Bias is everywhere, from top level press sources (New York Times/Washington Post), to academia(text books/papers), and most definitely the internet(forums and less credible press sources). Besides, I am planning to review/revise it so I do appreciate input and not only critcism on the bias that will never dissappear.
 
EbonySeraphim said:
...it's likely I'll determine it impossible to non-speculatively judge the performance of both consoles before the PS3 comes out and software is actually written for it altogether.
Exactly! ;)

If the bias supports false(or weakly true) reasoning and logic, you can disqualify it using correct(or stronger) logic or reasoning alone. If the bias supports correct(on the deepest levels of) logic and reasoning, then all you're getting mad at is the bias and the only thing left to do is flame the bias holder even if he is right. Or, people end up trying to argue against correct logic, with incorrect logic...is the cycle making sense?
Very true.

Besides, I am planning to review/revise it so I do appreciate input and not only critcism on the bias that will never dissappear.
Good for you. This is the best place I know of to ask questions about what you're unsure of, and get corrections (like ShootMyMonkey saying Xenos will be BW bound on some FB effects - I guess I ought to ask what effects he's talking and whether we're on the same wavelength in that regard!)
 
EbonySeraphim said:
Stop trying to call out my bias please. Bias is admittedly there so let's aim on making it better, not point out where it's wrong or biased. That is what leads to a flame war. Talk about the merits, not the bias. All people have bias no matter what they are talking about. If the bias supports false(or weakly true) reasoning and logic, you can disqualify it using correct(or stronger) logic or reasoning alone. If the bias supports correct(on the deepest levels of) logic and reasoning, then all you're getting mad at is the bias and the only thing left to do is flame the bias holder even if he is right. Or, people end up trying to argue against correct logic, with incorrect logic...is the cycle making sense?


Agreed & makes sense.

The problem with "bias" is when the un-informed read biased reports such as yours that are not only biased but also incorrect, it leads to false ideas about the subject. This is how we get people claiming ps3 is 10x more powerful than 360 and other such crazyness. Not that I'm saying you have quoted such a gross error in your report but un-informed reports which read as otherwise credible blur the line between truth and false and should be corrected where possible so as not to lead those who read the report to misconceptions about a subject or the author.

Also when this bias is shared and supported by a multitude of others in a similar fashion which also seem credible in their own right but are also misinformed and spreading this misinformation the results are even more impactful.

Overall I say good job sir and I look forward to the revised & balanced report! :smile:
 
I guess I ought to ask what effects he's talking and whether we're on the same wavelength in that regard!
Well, in simplest terms, it's generally a pain in the neck if you try and treat the Xenos like the GS and start going through framebuffer-wide blend operations or try to accumulate using backbuffer copies.

Accumulated lighting, for instance, over multiple passes is really death -- it's one of the things that really killed the framerate in at least 4 360 titles I can think of (which shall remain nameless). It is much cheaper to use texture arrays assigned to a single sampler register (for shadow maps) and just do all your lighting in one pass. Though it's fair to say that framebuffer bandwidth alone wouldn't be an issue here in practice, I can point to examples of texture-less demos where this still made nearly a 4:1 difference in framerate.

Of course, there's the more obvious case of MRT, since Xenos is (almost) in the same shoes as any other GPU in that case.
 
Last edited by a moderator:
Jesus2006 said:
Afaik all PS3 games have (or are confirmed to have) AA , WarHawk is 4xAA (@60fps), HS is 4xAA, Resistance i don't know but i guess at least 2xAA.

MotoGP does not have AA i think at least not the demo which i downloaded. The only 360 game i know to have 4xAA is DoA but this is "just" a fighting game with little geometry and not much on the scene.

My friend, you are very misinformed. The previous list of 360 games all have 2XAA-4xAA. If you only know DOA4, then you know very little. And to say that all PS3 games are confirmed to have AA is practically ignorance, just like Microsoft saying that 4xAA is "free".

As for the 360's AA capabailties, it gets better in time as developers find out about the hardware. For example, Forza 2, which will release in fall, will run in 60 fps, have 4xAA, 2XAA, and will feature better graphics than PGR3. This theory will no doubt apply to PS3 games as well.
 
Proelite said:
My friend, you are very misinformed. The previous list of 360 games all have 2XAA-4xAA. If you only know DOA4, then you know very little. And to say that all PS3 games are confirmed to have AA is practically ignorance, just like Microsoft saying that 4xAA is "free".

I think you didnt understand my point.

The PS3 has absolutely no hardware acceleration for MSAA and was said to have a too tight bus to efficiently render at 4xMSAA. Nevertheless, there are already games in development, which are confirmed (and shown at E3) to be 4xAA and these are launch games.

On the other hand there is Xbox360 which is known for its EDRAM which gives AA "for free" according to Microsoft. Still almost no launchgames had AA and most of the games still out right now do not have any AA as well ( i recently downloaded demos of Tomb raider, Burnout, Prey, Test Drive, PGR3... none of which has AA), some have 2xAA (Condemned, Oblivion, GRAW(only single player), Lost Planet).

I did some more web research and found that only 2 games on the 360 do really use 4xAA, this is DoA and Fight Night 3, both simple fighting games.
 
Jesus2006 said:
I think you didnt understand my point.

The PS3 has absolutely no hardware acceleration for MSAA and was said to have a too tight bus to efficiently render at 4xMSAA. Nevertheless, there are already games in development, which are confirmed (and shown at E3) to be 4xAA and these are launch games.

On the other hand there is Xbox360 which is known for its EDRAM which gives AA "for free" according to Microsoft. Still almost no launchgames had AA and most of the games still out right now do not have any AA as well ( i recently downloaded demos of Tomb raider, Burnout, Prey, Test Drive, PGR3... none of which has AA), some have 2xAA (Condemned, Oblivion, GRAW(only single player), Lost Planet).

I did some more web research and found that only 2 games on the 360 do really use 4xAA, this is DoA and Fight Night 3, both simple fighting games.


The RSX like every other NVIDIA chip for the last few years has built in hardware MSAA. I am still waiting for something in that is not prerendered on the PS3 that destroys all up coming 360 games.
 
quest55720 said:
The RSX like every other NVIDIA chip for the last few years has built in hardware MSAA.
Really? I''ve never heard of this one. Proof?

quest55720 said:
I am still waiting for something in that is not prerendered on the PS3 that destroys all up coming 360 games.

I am not sure of what are you trying to say here. To expect a game on any of the platforms to destroy the competition is wishful thinking. Given the little disparity between the 360 and ps3, it won't happen.
 
Proelite said:
Really? I''ve never heard of this one. Proof?



I am not sure of what are you trying to say here. To expect a game on any of the platforms to destroy the competition is wishful thinking. Given the little disparity between the 360 and ps3, it won't happen.


Acording to the author of the article the PS3 walks all over the 360 with out breaking a sweat. If one peice of hardware is so vastly superior it should have no problems destroying the compitition graphically. If the author is right where are the results or are all developers stupid?

So the RSX has no extra sampling units or extra compression when MSAA is used like ATI GPUs? If it does the it has hardware MSAA.
 
Jesus2006 said:
The PS3 has absolutely no hardware acceleration for MSAA and was said to have a too tight bus to efficiently render at 4xMSAA. Nevertheless, there are already games in development, which are confirmed (and shown at E3) to be 4xAA and these are launch games.
Once again, some, not ALL. I don't disagree with you on the former.

jesus2006 said:
On the other hand there is Xbox360 which is known for its EDRAM which gives AA "for free" according to Microsoft. Still almost no launchgames had AA and most of the games still out right now do not have any AA as well ( i recently downloaded demos of Tomb raider, Burnout, Prey, Test Drive, PGR3... none of which has AA), some have 2xAA (Condemned, Oblivion, GRAW(only single player), Lost Planet).

I did some more web research and found that only 2 games on the 360 do really use 4xAA, this is DoA and Fight Night 3, both simple fighting games.

Again, as I said that as developers gets to know how to work xenos, more games with 2xAA, and 4XAA will appear. Forza 2, once again, will be a game that allows for 720p, 60fps, 4xAA. Thus, the power of the EDRAM module is not "negligible."
 
quest55720 said:
So the RSX has no extra sampling units or extra compression when MSAA is used like ATI GPUs? If it does the it has hardware MSAA.

Nope, it doesn't. It's a N47 derivative. Btw, I don't think Xenos has hardware MSAA either. It has hardware tesselation, but not the former.
 
Last edited by a moderator:
scooby_dooby said:
Meaning when you have a new architecture that developers are not used to. You have to give them time to exploit that architecture before you can judge it's effectiveness. 8 months in is not nearly enough time to proclaim: " the theoretical features of Xenos are neglectible when it comes to real games"

Well in that case, you can use the early efforts to judge the amount of power that can be easily gained from the system, and compare platform optimized games later. Which looks better, Xbox 360 or Ps3 launch-window games?
 
EbonySeraphim said:
Stop trying to call out my bias please. Bias is admittedly there so let's aim on making it better, not point out where it's wrong or biased. That is what leads to a flame war. Talk about the merits, not the bias. All people have bias no matter what they are talking about. If the bias supports false(or weakly true) reasoning and logic, you can disqualify it using correct(or stronger) logic or reasoning alone. If the bias supports correct(on the deepest levels of) logic and reasoning, then all you're getting mad at is the bias and the only thing left to do is flame the bias holder even if he is right. Or, people end up trying to argue against correct logic, with incorrect logic...is the cycle making sense?

How can you make it better without pointing out what's wrong? Your entire statement, while sounding reasonable in theory, has not stood up in practice in any of the discussions I have seen. The problem with bias, is that it even when logical empirical evidence is presented to refute another point, bias can easily discredit, ignore, or argue such findings. Particularly in areas such as this, where clearly there are no right or wrong answers, and passions run so deep. Admitting bias and a lack of understanding of the compared architectures doesn't help to bolster the validity of your article. And contrary to your belief, there are those who can remain objective, and accurately and fairly present the facts whether they are personally bias or not. None of us are perfect, and the criticism was not and is not a personal attack. Good luck with the article.
 
Fox5 said:
Well in that case, you can use the early efforts to judge the amount of power that can be easily gained from the system, and compare platform optimized games later. Which looks better, Xbox 360 or Ps3 launch-window games?

PS3 all the way there :)
But really, there are already multiplatform games that are confirmed by the developers to look better on PS3 and this is while we are in 2nd(+) gen. Xbox 360 games and PS3 hasn't even launched (talking about CoD3, supposed to be a PS3 launchgame).

So it cannot be that hard after all to get something out of PS3 as has been said by some over and over again. In contrast, it seems pretty hard on 360, looking at the "AA-less" games/bad framerate titles etc.
 
@ quest55720, Jesus2006 & Proelite - this is a silly argument that's not really contributing to the thread.

@ Jesus2006 - XB360 doesn't have 'hardware acceleration' of MSAA. It has hardware that alleviates the BW demands of MSAA from the main memory BW. And it has FP16 and FP10 HDR MSAA.

quest55720 said:
The RSX like every other NVIDIA chip for the last few years has built in hardware MSAA.
Yes, they can do MSAA, but not on FP buffers used for HDR. Given bandwidth demands and lack of FP16 MSAA ability (assuming RSX hasn't got that), it is quite surprising the number of titles with AA on PS3 to date. And I can see Jesus2006's POV that Xenos unique contributions don't constitue any advantage. But Jesus2006 is not appreciating the eDRAM isn't just about AA, and that software for Xenos hasn't got it's head around the hardware yet. Also, if they are achieving the same effects as PS3, with less system bandwidth, Xenos is contributing. The two consoles have the same job to do, produce games with nice graphics, and both go about that in different ways. If both run UE3 with 4xAA, HDR, 60fps, what-have-you, their not better architectures, just different, both solving the same problems but in different ways. The eDRAM isn't about getting better graphics than PS3 neccessarily, but more about not impacting the rest of the system, allowing for the use of cheaper RAM, unified memory, and other design decisions. If XB360 doesn't outperform PS3 in the visuals, it doesn't at all mean Xenos isn't doing it's job.
 
Fox5 said:
Well in that case, you can use the early efforts to judge the amount of power that can be easily gained from the system, and compare platform optimized games later. Which looks better, Xbox 360 or Ps3 launch-window games?

While I agree that the 2006 crop of games looks better than the 2005 360 launch games, I don't see how it's very meaningful. First of all you have to ignore the full year of next gen development that developers have had since the launch of the 360, then you'd also have to ignore the fact that the PS3 launch was delayed 6 months.

With that said, I've always expected developers to get more out of RSX earlier, than with Xenos, due to the RSX being such a traditional architecture. 360 has a bigger learning curve on tne GPU, while PS3 has the biggest learning curve for the CPU, at least that's how I've always looked at it.

I've never expected Xenos's power to be 'easily gained' nor would I expect the same for CELL.
 
scooby_dooby said:
With that said, I've always expected developers to get more out of RSX earlier, than with Xenos, due to the RSX being such a traditional architecture. 360 has a bigger learning curve on tne GPU, while PS3 has the biggest learning curve for the CPU, at least that's how I've always looked at it..


But there is no learning curve. The unified shading is totally transparent to the developer. It's DX9 all the way. The problem is the EDRAM which is currently more of a bottleneck than a feature for most developers.

And this is just something the PS3 has not to struggle with.
 
Jesus2006 said:
But there is no learning curve. The unified shading is totally transparent to the developer. It's DX9 all the way. The problem is the EDRAM which is currently more of a bottleneck than a feature for most developers.

And this is just something the PS3 has not to struggle with.

How about you and me both shut up for this thread because neither of us knows enough about the respective system's architecture compared to the senior members. We're just derailing the thread.
 
Status
Not open for further replies.
Back
Top