Alternative AA methods and their comparison with traditional MSAA*

You see,i just dont believe that.Dont get me wrong,i love SM but this is purest example of PR since the ad when they said cell is maybe smarter then humans.

Anybody who knows anything about tech side of consoles will just pass this PR bull but some fanboys will stick to this and "dream" on about next SM title looking twice as good as GOW III...

I don't know if anyone would argue that the jump from Uncharted 1 to Uncharted 2 is just a PR talk (they also gave a conflict statement either: 30% and sometimes they said it might be misleading to give a percentage). So it is all in our head to use inductive reasoning to figure it out. I would not jump and accuse somebody a fanboy when I don't think it through.
 
Hah good call. It just amazes me how devs internally contradicts each other or themselves to deliver some hype for a nice check, else why would they?

To bad the limit will never be reached until the 95-99% bar swings over to 100%. Then full GI with raytracing will be achievable and tesselation on all surfaces. at 60fps. :D

Are they really contradicting ? The GT video says GoW3 used all the processors/cores, but they didn't say they use up 100% of every cores. There was another interview where the devs admitted that GoW3 does not stress the hardware as much as U2/KZ2. The average % utilization of all the SPUs could be lower than 50%.

I think IGN said KZ2 used 60%.

Anyway, these are very rough figures. Without understanding the exact technical context, it's hard to make any sense out of them. They are claiming that there is more headroom for growth.
 
Hah good call. It just amazes me how devs internally contradicts each other or themselves to deliver some hype for a nice check, else why would they?

To bad the limit will never be reached until the 95-99% bar swings over to 100%. Then full GI with raytracing will be achievable and tesselation on all surfaces. at 60fps. :D



Define "last minute". Also scheduling scheme to improve perfomance 8% sounds kinda moot if there would have been so much "untapped potentials" after ~4 years of development.

Scheduling is just one of many problems they have. Imagine if they can improve most of their algorithms?

by the way, here is the definition of last minutes:
@ChristinaCoffin Random Scheduling Opt courtesy of Cedric in last week of project = +8%. Went in after first review code lol.
http://twitter.com/TDMoss/status/9854078121
 
Last edited by a moderator:
I don't know if anyone would argue that the jump from Uncharted 1 to Uncharted 2 is just a PR talk (they also gave a conflict statement either: 30% and sometimes they said it might be misleading to give a percentage). So it is all in our head to use inductive reasoning to figure it out. I would not jump and accuse somebody a fanboy when I don't think it through.

Ok so you are telling me that you find Uncharted 2 looking 3 times better then UC1?Because they said its 30% with UC1,considering that UC2 was their second game on ps3,thus they probably knew what to do and what no to with hardware,but they still said they use 100% of power.But i would not be sure that UC2 looks that much better then its Uncharted 1.Sure its never optimized enough but its not like UC3 will suddenly look alot better...

Its all rough estimates,which are not really true and i dont think anyone would argue with me if i said that if they choose to develop GOW III again they would not get much better results.
 
I think IGN said KZ2 used 60%.

And U2 was 100% that later got changed to 80%.
Anyway, these are very rough figures. Without understanding the exact technical context, it's hard to make any sense out of them.

True, could be avg, minimum, max or if even correct at all and not hype marketing etc.
 
And U2 was 100% that later got changed to 80%.

Yea you are right.It seems like when game is just to be released it uses 100% and when some time passes it suddenly is just not really optimized thus next tittle will look alot better.Those kind of things just give a fanboys false hope...
 
May want to include the links or interviews for the context. Quoting numbers out of the blue is seldom a good idea.

Both of you seem to have mixed up "processors" and "PS3 power" above. In the U2 case, 100% of what compared to 80% of what ? [confused]
 
Ok so you are telling me that you find Uncharted 2 looking 3 times better then UC1?Because they said its 30% with UC1,considering that UC2 was their second game on ps3,thus they probably knew what to do and what no to with hardware,but they still said they use 100% of power.But i would not be sure that UC2 looks that much better then its Uncharted 1.Sure its never optimized enough but its not like UC3 will suddenly look alot better...

Its all rough estimates,which are not really true and i dont think anyone would argue with me if i said that if they choose to develop GOW III again they would not get much better results.

you can't just use elementary calculations to judge quality. But you can't deny the significant jump in quality from UC1 to UC2. That is all that matters.
GoW3 team gave you the 50% efficiency figure, that would give an estimate of improvements in quality of the next game.

Yea you are right.It seems like when game is just to be released it uses 100% and when some time passes it suddenly is just not really optimized thus next tittle will look alot better.Those kind of things just give a fanboys false hope...

I think your argument has no ground. They proved their number by showing the jump in quality from UC1 to UC2.

What I can tell you is this: if Naughty Dog uses MLAA in their next game, they would get 5ms back from the GPU and get a much better AA.
So if you use elementary calculations 5/33 = 15% improvements and that is not all they can improve.
 
Last edited by a moderator:
Both of you seem to have mixed up "processors" and "PS3 power" above. In the U2 case, 100% of what compared to 80% of what ?

I am talking stritly CPU. For numbers just google it up and get lots of results. Or just browse the U2 thread, I am sure you already read it before. But I'll leave it there. :smile:
 
And U2 was 100% that later got changed to 80%.

Yea you are right.It seems like when game is just to be released it uses 100% and when some time passes it suddenly is just not really optimized thus next tittle will look alot better.Those kind of things just give a fanboys false hope...

I think your argument has no ground. They proved their number by showing the jump in quality from UC1 to UC2.

I am done with discussion but why are you false quoting me?

The second line I have not written.
 
I am talking stritly CPU. Numbers out of the blue, hah. Just google it up and get lots of results. Or just browse the U2 thread, I am sure you already read it before. But I'll leave it there. :smile:

I did a quick google...

The 80% PS3 power was mentioned by Kazunori when he talked about GT5:
We have an interesting article for you to read now, as Gran Turismo chief Kazunori Yamauchi has admitted that Gran Turismo 5 does not use all of the PS3’s processing power, even adding that it uses roughly the same amount as seen in Uncharted 2.

As reported from Eurogamer, Yamauchi spoke in a recent interview saying that only 80% of the PS3’s total power has been used for the upcoming racer, reminding everyone that GT5 has taken 5 years to make and that nothing is getting easier or less expensive.

... and it's still not out yet. >_<
By the time it's released, may be it will max out PS3 like U2 ?


Naughty Dog mentioned they used close to 100% PS3 "power" themselves:
http://www.vg247.com/2009/10/23/uncharted-2-used-close-to-100-of-ps3s-power/

Naughty Dog co-boss Christophe Balestra’s claimed Uncharted 2 almost maxes out PS3’s “power”.

“On PlayStation 3 for our first game, we were using probably only 30 percent of the power of the Cell, and this time we’re getting close to 100 percent of using the power of the PS3,” he said, talking in two making-of movies you can see after the break.


All these are interesting but very rough estimates. However, I don't see ND contradicting themselves ? unless you can find the 80% CPU utilization quote from NaughtyDog ? I did find an earlier ND interview that predicted that U2 would use close to 100% of PS3 power.

Note that the developers may get better as they gain more experiences. Back on topic, we know the GoW3 team is also interested in 3D vision. So there should be some spare resources lying around.
 
It certainly could. No question. Hypothetically, if you make it lower quality in exchange for a performance boost, that is still going beyond the scope of the original implementation.
If you make it run in javascript, that's going beyond the original implementation.

Etc. All we know is their implementation differs from the standard method.



Yes, my understanding is they use temporal AA (where you swap the AA pattern in alternate frames). It's apparently controlled by the 'reduce flickering' option in their menus.

I believe that would be a different implementation or deviation from the original, but not "beyond" it. It would have to, at least, be better, as a whole, than the original implementation (to qualify as "beyond"). What are the chances of us seeing their exact implementation for determination?
 
50% of power?O really?I remember he said one time that it uses all of the cpu power...here

http://psinsider.e-mpire.com/index.php?categoryid=17&m_articles_articleid=1247

Not only that but the way they said it would never be able to run on 360 and now saying its only 50 % of ps3 power would make you think that ps3 is more then twice as powerful as 360...

Oh my God what PR and hype does...they should just admit,we are using it to its limits,nobody would that that for bad,but this...he will trick alot of fanboys with this PR nonsense.
WTH of post is it? & Why someone even reply? First: it's OT. Second: SM talking of GOW engine & it use of 50 % of CPU; third, said this game it only possible on the ps3 not means automatically 50% of ps3 so double powerish of the rival. Before to talk of PR nonsense learn to read the interview.
 
Last edited by a moderator:
I believe that would be a different implementation or deviation from the original, but not "beyond" it. It would have to, at least, be better, as a whole, than the original implementation (to qualify as "beyond"). What are the chances of us seeing their exact implementation for determination?

Why is it not beyond when there is no drawback that is from the original implementation?
 
I think IGN said KZ2 used 60%.

If I remember correctly, KZ2 wasn't fully optimised. I'm sure that I've seen a presentation from Guerilla where they say if it works, leave it alone - don't bother to optimise it. That's the key isn't it? You could keep all of the cores busy all of the time, but it doesn't mean that they are being used to their optimum level. Thus it makes such figures rather pointless.

for( ; ; )
{
}

Job done! ;)
 
I am talking stritly CPU. For numbers just google it up and get lots of results. Or just browse the U2 thread, I am sure you already read it before. But I'll leave it there. :smile:

If you are strictly talking about CPU usage, why are you piggybacking on Ruskie's post? He was clearly talking about power and not just CPU usage.

Anyway, ND said they were probably going to use CLOSE to 100% of the CPU cycles. They, also, said that you have not seen the last improvements for their engine. ND said they just got "all the plates spinning". Now, "we can make them spin faster" (I recalled that from the G4TV interview last year). They said their systems weren't really optimized. They were just ported over to the SPUs. It may be possible for ND to get a 50% to 100% increase in overall speed of their code. Only time will tell.

BTW, I thought GoW3 used 5 SPUs from last year's presentation. Can someone confirm or deny?

EDIT: Here is the U2 interview. http://g4tv.com/thefeed/blog/post/6...lls-and-amy-hennig-uncharted-2-interview.html
 
Last edited by a moderator:
Back
Top