By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Sales Discussion - is the x-box 720 DOA?

HappySqurriel said:

Please, never work on a project I am working on!

You seem to have very little understanding of the difference between PR bullshit and what really matters. Having built several raytracers when I was in university, I can tell you that Raytracing performance is largely dependant on the choices of intersection tests you choose and the datastructures you use to limit the number of intersection tests you have to perform. The Cell processor doesn't offer much in the way of a per-clockcycle improvement in the intersection tests performance, and offers little to improve the performance in managing the datastructures; it does offer parallel processing of the intersection tests but the performance gains will mostly be limited by the loss caused by load balancing across all the threads.

In other words, the only reason they can make a claim like this is because Intel, AMD and IBM haven't made a single core processor in several years ...

  Sure, I will make sure we never work on a project together.

 Intel and AMD still make single core CPU's

Also, go here, and read a bit.

http://www.ibm.com/developerworks/power/cell/index.html?S_TACT=105AGX16&S_CMP=LP

This is a nice paper on ray-tracing on the Cell.

http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/05CB9A9C5794A5A8872570AB005C801F/$file/2056_IBM_TRE.pdf

A good read, but for the ppl not interested in reading, here is a part of the paper:

 

For benchmarking purposes the following values 

were selected: 

 1280x720 (720p) output image size 

 7455x8005 Map size 

 2048 map steps to full haze 

 1.33 x (2 – 8 Dynamic) or ~2-32 samples per pixel Visualization and Computer Animation 

 

With these settings the following rendered frame 

rates were measured: 

 

System / frames per second 

2.0 GHz Apple G5 1GB memory (no image encode)  0.6 

3.2 GHz UP Cell 512MB memory 30 

3.2 GHz 2-way SMP Cell 1GB memory 58 

  



Around the Network
Viper1 said:
Realmofo, the CELL is one of the best chips currently at many "specific" tasks. Put it against my Core 2 Duo for many general purpose tasks and my C2D will wipe the floor with it.

You cannot just assume that because the CELL currently leads in certain tasks that its product line will always do so. That's a preposterous notion in the world of chip technology.


 Oh one day the Cell will be nothing compared to other CPUs. I am just assuming that is not going to happen in the next 2-3 years, and that's all the time MS has to come up with an affordable CPU for the next xbox.

Also, yes, as certain tasks, the C2D in the Mac Book Pro I am writing this on is faster then the Cell. But it's only maybe twice as fast. At the things the Cell is faster at, it's orders of magnitudes faster. Put all that together, and smart game developers will make games with much more of a "wow" factor for gamers.

They will just take advantage of what the Cell does well, and make things we have never seen before.



TheRealMafoo said:
HappySqurriel said:

Please, never work on a project I am working on!

You seem to have very little understanding of the difference between PR bullshit and what really matters. Having built several raytracers when I was in university, I can tell you that Raytracing performance is largely dependant on the choices of intersection tests you choose and the datastructures you use to limit the number of intersection tests you have to perform. The Cell processor doesn't offer much in the way of a per-clockcycle improvement in the intersection tests performance, and offers little to improve the performance in managing the datastructures; it does offer parallel processing of the intersection tests but the performance gains will mostly be limited by the loss caused by load balancing across all the threads.

In other words, the only reason they can make a claim like this is because Intel, AMD and IBM haven't made a single core processor in several years ...

  Sure, I will make sure we never work on a project together.

 Intel and AMD still make single core CPU's

Also, go here, and read a bit.

http://www.ibm.com/developerworks/power/cell/index.html?S_TACT=105AGX16&S_CMP=LP

This is a nice paper on ray-tracing on the Cell.

http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/05CB9A9C5794A5A8872570AB005C801F/$file/2056_IBM_TRE.pdf

A good read, but for the ppl not interested in reading, here is a part of the paper:

 

For benchmarking purposes the following values 

were selected: 

 1280x720 (720p) output image size 

 7455x8005 Map size 

 2048 map steps to full haze 

 1.33 x (2 – 8 Dynamic) or ~2-32 samples per pixel Visualization and Computer Animation 

 

With these settings the following rendered frame 

rates were measured: 

 

System / frames per second 

2.0 GHz Apple G5 1GB memory (no image encode)  0.6 

3.2 GHz UP Cell 512MB memory 30 

3.2 GHz 2-way SMP Cell 1GB memory 58 

  


So, to prove your point about Ray-TRACING you link to a paper about Ray-CASTING ... If you didn't realize this, Ray-Casting is the Technique that John Carmack used to produce 3D graphics with Doom; it is quite a bit different than Ray-Tracing. The G5 was released in 2002 and for its entire life was considered to be a worse performing processor than the equlivalent Pentium 4 and Athlon processors; the Cell must be uber, it outperforms a processor that is 5 years older than it and was considered mediocre when it was released.

You should really be quiet if you truely don't know what you're talking about!



HappySqurriel said:

So, to prove your point about Ray-TRACING you link to a paper about Ray-CASTING ... If you didn't realize this, Ray-Casting is the Technique that John Carmack used to produce 3D graphics with Doom; it is quite a bit different than Ray-Tracing. The G5 was released in 2002 and for its entire life was considered to be a worse performing processor than the equlivalent Pentium 4 and Athlon processors; the Cell must be uber, it outperforms a processor that is 5 years older than it and was considered mediocre when it was released.

You should really be quiet if you truely don't know what you're talking about!


I sent you a link to a ton of info on the Cell. If you had bothered to look around, you would have come across this:

http://www.alphaworks.ibm.com/tech/irt 

But that would require you to actually care about learning, and discoivering what the Cell can do. Instead, you would rather insult me (twice now).

You did give me good advice though. As it's obvious I can not have a civil conversion where in the end we might both learn something, because you seem to know it all, I will stop talking now. 



Really... People go off their heads trying to prove that the cell is a superior proccessor, HappySqurriel is making great points against it.

Which gaming developer wates to spread gaming codes over 8 cores anyway?



Current Consoles: Xbox 360 Elite, Playstation 2, Gaming Rig, Nintendo Wii, Playstation 3.

Xbox Live: Jessman_Aus - Playing: Ace Combat 6, Fifa 09

Playstation Network: Jessman_Aus - Playing: MGS4, Resistance 2

Wii Freind Code: 3513-9191-8534-3866 - Playing: SSBB

Brawl Code: 1590-6125-1250

Xfire: J3ssman - Playing: Fallout 3, Farcry 2

Jessman: Fears the Mangina

 

                                

Around the Network
Jessman said:
Really... People go off their heads trying to prove that the cell is a superior proccessor, HappySqurriel is making great points against it.

Which gaming developer wates to spread gaming codes over 8 cores anyway?

That's the direction we're headed Cell or no Cell. Newer PC cpus like Intel Core Duos are quad cores, and we definitely won't be stopping there.  64-core CPUs are already in production.



Since sony dropped support of cell processor development, it is almost certain that the next Playstation game console will not use it.



I don't need your console war.
It feeds the rich while it buries the poor.
You're power hungry, spinnin' stories, and bein' graphics whores.
I don't need your console war.

NO NO, NO NO NO.

Shame. They spent a whole bunch of money on it; it would have been nice to use some aspects of it and not have to re-invent the wheel as it were.



Wow, RealMafoo, you post links to IBM's website to help back up your claims about the Cell? Nice...
Happy, I really feel for you...this is almost like the old convos, about the cell and PS3, between MikeB and I can't remember his name. It has to be awfully frustrating dealing with people like that, who ignore factual reasoning and intelligent arguments, shooting your posts down with IBM PR and information on something different then what is being argued. The whole G5 thing really gave me a laugh, I mean...come on? Are you serious??!?!?



Currently Playing:

PS4 - Killzone:SF and Assasins Creed 4

 

XBox One: BF4, CoD:Ghosts, Dead Rising 3, Forza 5

 

Changing channels with my voice: priceless!!!

WEWdeadeye said:
Wow, RealMafoo, you post links to IBM's website to help back up your claims about the Cell? Nice...
Happy, I really feel for you...this is almost like the old convos, about the cell and PS3, between MikeB and I can't remember his name. It has to be awfully frustrating dealing with people like that, who ignore factual reasoning and intelligent arguments, shooting your posts down with IBM PR and information on something different then what is being argued. The whole G5 thing really gave me a laugh, I mean...come on? Are you serious??!?!?

 Sigh, you guys pick an argument, and no matter how much code gets written to disprove your views, you dismiss it as PR trash.

Here are some non-IBM links:

 http://portal.acm.org/citation.cfm?id=1229445

"Our experiments show that the Cell benefits significantly from dynamic methods that selectively exploit the layers of parallelism in the system, in response to workload fluctuation. Our scheduler out performs the MPI version of RAxML, scheduled by the Linux kernel, by up to a factor of 2.6. We are able to execute RAxMLon one Cell four times faster than on a dual-processor system with Hyperthreaded Xeon processors, and 5--10% faster than on a single-processor system with a dual-core, quad-thread IBM Power5processor.

 

http://www.citeulike.org/user/Zakalwe/article/1243198

"using a combination of low-level optimized kernel routines, a streaming software architecture, explicit caching, and a virtual software-hyperthreading approach to hide DMA latencies, we achieve for a single cell a pure ray tracing performance of nearly one order of magnitude over that achieved by a commodity CPU

These people don't care about some stupid PS3 vs 360 argument. They are not using the CELL to prove a point, or win an argument. They are scientists getting real world results out of a very powerfull processor.

I guess you guys want to believe the CELL sucks, so it probably wouldn't matter if someone could do something like crack passwords 100 times faster then ever before, you still wouldn't believe it holds any more power then the 360's CPU. It's like a religion to you people. I am done.

P.S. It been done:

 http://www.pcworld.com/article/id,140064-c,gameconsoles/article.html