CELL PROCESSOR "locked SPE" Question.....

[Persistantthug]

Hi...
I'm new, and I don't really have any great deal of technical knowledge....well, not like you guys of course, but I have a question, if you don't mind.

My question is, about the PS3's Cell Processor's SPE's.
As I'm aware, been told rather, the Cell Processor has 8 SPE's, 2 of which are not used for gaming because 1 of the 2 is used for the PS3's OS and some other protocols, while the other is idle or "locked out" because early on Sony may have been having trouble getting all 8 to work on a significant enough amount of CELLS...So basically, "7's good enough", if you will.

My question is, what would it take for Sony Developers to tap in and unlock this 8th and idle SPE for use in games without hurting PS3's and/or causing "bricked" units or whatnot?

Do you think it's likely and or somehow possible for future titles down the road, say, 2 years from now or somthing like that?
Thoughts?

Thank you.

 

[AmFreak]

The 8th isn't used cause of the higher yield. By only using 7 SPU's they could also use the CELL's with one broken SPU. That's why they never will use the 8th one, cause there are PS3's out there which haven't a working one.

 

[Silent_Buddha]

Aye, if some developer were able to somehow unlock the 8th SPU for use in their games, then that game wouldn't work on any PS3's where the 8th SPU truly is non-functioning.

Short answer, never going to happen.

Regards,
SB

 

[Persistantthug]

Aye, if some developer were able to somehow unlock the 8th SPU for use in their games, then that game wouldn't work on any PS3's where the 8th SPU truly is non-functioning.

Short answer, never going to happen.

Regards,
SB

You know what?
I'm guilty of not asking the right question(s).

You see, I know there are a certain (unknown) number of early PS3's with "defective" (only 7 SPE's work) CELL PROCESSORS.
However, the heart of my question....let me stop beating around the bush...

 

[Persistantthug]

With the 8th SPE.....

With the 8th SPE working (hypothetically)....

Would or could 1080p @ 60 frames per second be possible, with no quality degradation in GOD OF WAR 3 and UNCHARTED 2?


That's my true question for the experts here.
Thank you.

 

[draconian]

Why does it matter? No one can use that 8th spu. And no, one more spu isn't going to do anything.

 

[Jedi2016]

With the 8th SPE working (hypothetically)....

Would or could 1080p @ 60 frames per second be possible, with no quality degradation in GOD OF WAR 3 and UNCHARTED 2?


That's my true question for the experts here.
Thank you.

There are a lot of other things to consider with 1080p/60 than just processor power, and the tiny boost given by one more SPE wouldn't make much difference in the long run.

I don't think the SPEs themselves really help much when it comes to graphics. Although I don't have much in-depth knowledge of the system's workings, I was under the impression the SPEs were used for things like physics and AI subroutines.. basically helping to run the underlying code of the game rather than contributing to what's being displayed graphically. Even if they were used for things like computing lighting and shading, the only result of an extra SPU would be a slight improvement in the overall quality of the image (and by "slight", I mean "most people wouldn't even notice"), rather than a boost in resolution or frame rate. The GPU is the heart of that part of it.

 

[flynn]

There are a lot of other things to consider with 1080p/60 than just processor power, and the tiny boost given by one more SPE wouldn't make much difference in the long run.

I don't think the SPEs themselves really help much when it comes to graphics.

Actually the SPEs play a huge role in graphics in many games, Uncharted 2 being the obvious example, but you are correct in saying that it takes a lot more than an extra SPE to bump up resolution. Frame buffer sizes change which will most like impact your memory allocation strategy, the GPU has more work to do, etc.

 

[Laa-Yosh]

This thread is silly.

 

[Jedi2016]

Actually the SPEs play a huge role in graphics in many games, Uncharted 2 being the obvious example, but you are correct in saying that it takes a lot more than an extra SPE to bump up resolution. Frame buffer sizes change which will most like impact your memory allocation strategy, the GPU has more work to do, etc.

I would imagine that the SPEs would be used to process certain graphical elements, maybe? Things like shaders and lighting that requires more advanced computation to actually figure it out in addition to displaying the result? (I'm asking, I honestly don't know. )

Most of my knowledge of these sort of things come from my work in CGI. In that world, there are several things that are processed ahead of the actual render, including radiosity, subsurface scattering, etc. This information is then fed into the final render, and that render takes about the same amount of time as it normally would without those effects.

For example, let's say you have an interior scene that renders in ten seconds without radiosity. If you add it, it might take the computer five minutes to compute the bouncing light rays, but when it feeds this into the render engine, the actual drawing of the render still takes only ten seconds. Is that the sort of thing that could be offloaded to the SPEs?

That would go back to what I theorized earlier, that the addition of more processing power might make for a small improvement in the quality of the image, in terms of advanced shading or whatever, but wouldn't be able to make any real contribution to bumping up the resolution.

Again, going back to the CGI example.. the larger an image, the longer it takes to render, and it's usually equal to whatever the increase in pixel count is. So, let's say a 1280x720 image takes ten seconds to render. That's 921,600 pixels in ten seconds. A 1920x1080 image is 2,073,600 pixels.. more than double the 720p image. And, in truth, it takes roughly twice as long to render a 1080p image as it does a 720p image, without factoring in any pre-render computation like radiosity. I can do some tests if anyone wants to see some actual numbers. Granted, a game behaves in a very different fashion, where it renders a 720p image in a small fraction of a second. But the basic logic should be much the same.. twice the pixels equals twice the drawing time (but not necessarily twice the computational time).

Going back to the original argument (as pointlessly hypothetical as it is), the tiny boost from a single SPU (really, quite tiny, I'd imagine.. like overclocking a processor by 2% or something) simply isn't going to make any appreciable difference in the system's performance, especially to the average player.

 

[Arwin]

The answer to the question is a big, fat, NO. End of thread.

If you're wondering why, imagine how much more work rendering 2x the amount of detail requires. Now think about how much more performance you are getting from just one additional SPU in the system (roughly an increase of 1/8th of CPU power). Finally, consider how much the CPU matters in the greater whole of RSX, memory bandwidth, etc.

 

[Persistantthug]

Going back to the original argument (as pointlessly hypothetical as it is), the tiny boost from a single SPU (really, quite tiny, I'd imagine.. like overclocking a processor by 2% or something) simply isn't going to make any appreciable difference in the system's performance, especially to the average player.

I'm a complete tech noob, pretty much...

But I can almost guarantee on the little knowledge that I do have,
That bringing in another SPE is more than the 2% you've estimated...especially since parallel is in the mix here.

Again....me...noob,

But I bet it's in the neighborhood of 10 to 15 percent...give or take 5.

The rest makes sense though.

 

[Ruskie]

I'm a complete tech noob, pretty much...

But I can almost guarantee on the little knowledge that I do have,
That bringing in another SPE is more than the 2% you've estimated...especially since parallel is in the mix here.

Again....me...noob,

But I bet it's in the neighborhood of 10 to 15 percent...give or take 5.

The rest makes sense though.

8th SPU will never get unlocked because if it was,then the game that would use 8th SPU would not work on alot of ps3 that got one of them not working and that would be extremely stupid business decision if anyone would even do that.In fact it would go down in the books.Anyway,that mere boost in performance is not even worth of dreaming of unlocking it

As for 8th SPU bumping up resolution thats also never going to happen because that "thing" depends more on systems bandwidth and gpu than its processing power.

 

[Svensk Viking]

I'm a complete tech noob, pretty much...

But I can almost guarantee on the little knowledge that I do have,
That bringing in another SPE is more than the 2% you've estimated...especially since parallel is in the mix here.

Again....me...noob,

But I bet it's in the neighborhood of 10 to 15 percent...give or take 5.

The rest makes sense though.

I am also a noob...But I have learned that we noobs shouldn't try to come up with assumptions regarding things we don't know, especially when talking with those who actually work with this

 

[Persistantthug]

As far as unlocking,

Couldn't it be done "situationaly"....say, if NAUGHTY DOG wanted the extra boost to get UNCHARTED 3 at 60 frames (again hypothetically), couldn't they create code to run a check to see if all 8 SPE's are working, and if they are, then Uncharted 3 will run at 60 frames....if not, then the game reverts back to 30 frames. Something of a "toggle on...toggle off...sorta.

So basically, the only developers able to take advantage of this are 1st party developers and others who have permission from Sony and only on specific games for specific purposes in order to get more "juice" if you will.


Is this possible? Why/why not?

 

[Neb]

As far as unlocking,

Couldn't it be done "situationaly"....say, if NAUGHTY DOG wanted the extra boost to get UNCHARTED 3 at 60 frames (again hypothetically), couldn't they create code to run a check to see if all 8 SPE's are working, and if they are, then Uncharted 3 will run at 60 frames....if not, then the game reverts back to 30 frames. Something of a "toggle on...toggle off...sorta.

A SPE would not be remotely near to deliver twice the perfomance. Then the thing about those with PS3s that wont get any "boost" getting dissapointed. It would be a lottery. And not sure if Cells with 8 good SPEs are even placed in PS3s, they might be for workstation products.

I'll bet the graphic chips also have disabled shaders to up yields.

 

[Akumajou]

I think that the only realistic way to answer your question would be to take a time machine and change history to make Sony wait to launch PS3 until 2009 when they could have IBM manufacture CellBE on a 45nm process and most likely a GPU design from nvidia (G92 based unless there would be a miracle to bring GF285) at 55nm (40nm was possible this year then again thats up to debate) and contemporary technology like a faster BR drive, more XDRDram and GDDR (most likely two times 512MB )...

The only problem would be that it would take up to two plus years to see Uncharted 2...

Additionally the competition would pretty much be able to fully establish a massive user base, developers would or might shift to another console, piracy would ravage the living crap out of the PS2 software sales much like it did leading up to PS3 launch, the fickle so called hardcore gamers would still criticize the price while ignoring the technology and capabilities like BC, the economic melty blood would definetly not help, being that other consoles would be established, some third parties would be too complacent with mature "easy to develop tools", Sony shareholders would be pulling their hair at the thought of letting the competition come out early, the blu ray vs hddvd format war would have dragged longer, costing more than it did and most important of all the phrase diminishing returns would rear its ugly head in the form of the games not really looking that different other than being in a higher resolution and framerate that average joe blow would take for granted and ignore.

I believe unless I am mistaken that Sony's top engineers highly desired to launch PS3 at 65nm to reach frequency goals and help keep thermals in check but realisticly 65nm was more than two years away in 2005 and even though 90nm was there there was no way to avoid the competition launching early like it did and Sony had to make sacrifices because back then the expectation and assumption was that the hardcore gamer, mostly in the USA and else where were going to eclipse the hunger frenzy for the PS2 launch in 2000 so a decision had to be made on the fabrication process of Cell and so one SPE is redundant however if Sone were to have known that demand for PS3 was going to be lower they could have cherry picked CellBEs to run at 3.2Ghz (as well as cherry picked RSX) and have all
SPEs enabled, but the problem with thinking of that being possible reguardless of the process being mature is that yields would dictate that there would be alot of CellBEs that would not qualify for PS3 and their clock speed would have to be lower and go else where and that gets really expensive.

Its physically impossible to assume that the 8th SPE could be enabled because of this, it would definetly make a difference yes but the cost for doing so is staggering when you really think about it...mainly because once the decision is made of a console spec for mass production there is nothing you can do, Sony kind of had a smooth transition with PS2 as prototype boards had .250nm EE and GS chips that required fans blowing air on top of heatsink while the final mass production has .180nm EE GS that was able to get by with the heatsink design the console had but Sony could have easily cranked up frequency on EE to 400Mhz or more and GS at 200Mhz or more as well as include twice the ram but then it would have been hard to avoid that the PS2 would have stayed at minimum $500 for at least two years.

Its possible also that microsoft would just replace the 360 in case gamers would have really gone into a frenzy on a years later launched PS3, but thats unknown.

I suggest you try to hunt down the subtitled The making of Killzone 2 to get a really good idea on how much of an impact the SPEs make on game development, there are a couple of videos on uncharted 2, the one called mastering the cell processor. Its very informative.

I think its a good question though but the reality of the decisions leading up to yields were made way back in 2004 or earlier as they projected 90nm production scenarios.

 

[Shifty Geezer]

But I can almost guarantee on the little knowledge that I do have,
That bringing in another SPE is more than the 2% you've estimated...especially since parallel is in the mix here.

Too many processing units with not enough local memory (cache) to support them sees them sitting idle, so just being parallel isn't any guarantee of acceleration. However, because the SPEs are self-contained, as long as the system bandwidth isn't being saturated, you'll get of the order of a whole unit speed improvement for each added SPE, meaning 1/7th increase in case of adding an 8th SPE to 7. This is not enough to make a huge difference to a game's visuals, certainly the difference from 30 to 60fps!

Is this possible? Why/why not?

Firstly we don't know if the extra SPE is being disabled in hardware or software. Secondly, how do you think naive Joe Public would feel if they found out with Uncharted 3 that their particular PS3 wasn't as good as his mate's? The public isn't informed enough, and never will be, to be able to deal with randomly more capable PS3s. Every buyer expects their PS3 to work the same as everyone else'sk, and rightly so.

And not sure if Cells with 8 good SPEs are even placed in PS3s, they might be for workstation products.

By now I expect the vast majority of Cells fabricated will be fully functional. It's a well matured chip!

 

[Shifty Geezer]

I think that the only realistic way to answer your question...

the realistic answer has already been stated - "no."
No to unlocking the SPE, no to the public accepting it, and no to it making a significant difference to the capabilities of extra-SPE PS3s (other than, with every game being patched, dedicating this SPE to GWAA for mich improved IQ on a lot fo games).

 

[Akumajou]

the realistic answer has already been stated - "no."
No to unlocking the SPE, no to the public accepting it, and no to it making a significant difference to the capabilities of extra-SPE PS3s (other than, with every game being patched, dedicating this SPE to GWAA for mich improved IQ on a lot fo games).

ok but I still believe that the engineering process had alot to do with the decision of the redundant SPE making it impossible and therefore a no answer. but I understand what you mean.