Starbreeze take on the Ps3 vs Xbox 360 (the Darkness Int)

[GB123]

CELL is superior on:

- total amount of internal memory (2304 KB versus 360's 1024 KB: 2.25 times)
- number of processors (8 versus 360's 3: 2.7 times )
- number of *concurrent* threads (8 versus 360's 3: 2.66 times)
- GFLOPS rating (204 GFLOPS versus 360's 77 GFLOPS: 2.65 times)
- integer rating (~51.2 billion instructions per second versus 19.2 BIPS: 2.66 times)
- internal bandwidth (204.8 GB/s versus shared 102 GB/sec L2 cache bandwidth: ~2 times)
- external bandwidth (60 GB/sec versus 360's 22 GB/sec: ~3 times)

On the order of 2 to 3 times for everyone of those specs/features over the Xbox 360 CPU. The only counter argument over all these months is that it will be harder to program, and that is agreed apon. Some of those numbers might be off by a bit, but certainly not by a lot, and no amount of explaination can make them go away.

It's hard to argue with facts, and that's not "myth", "legend" or "PR". But keep it up, everyone loves to hear again how superior CELL is over Xenon.

Thats just laughable..

If the PS3 is this much more powerful than the 360 why is the first thread of this post saying that the 360 and PS3 will have barely any visual difference.

This is not the only Dev who has said this.

By your way of thinking the PS3 will have a 2 x performance of total system resources than the 360, and that would be easily visable to developers.

The main reason many developer won't go on record to say one system is better than the other is because they don't have the experience to make an accurate judgement.

Seems to me the only dev's saying anything are those that are exclusive to the platform and it is in their best interest to big up that console.

 

[Laa-Yosh]

It's hard to argue with facts, and that's not "myth", "legend" or "PR". But keep it up, everyone loves to hear again how superior CELL is over Xenon.

Damn it, stop with the stupid PS3 propaganda stuff already! There's a lot of interesting discussion going on nowadays but all the time I have to face your stupid posts, it's really getting on my nerves now...

 

[AlNom]

If the PS3 is this much more powerful than the 360 why is the first thread of this post saying that the 360 and PS3 will have barely any visual difference.

To be fair, the Starbreeze guy was referring to the screenshots. Frame-rates may differ wildly (or not).

The main reason many developer won't go on record to say one system is better than the other is because they don't have the experience to make an accurate judgement.

or as multiplatform developers, they choose to be diplomatic about their words.

 

[Dr. Nick]

Damn it, stop with the stupid PS3 propaganda stuff already! There's a lot of interesting discussion going on nowadays but all the time I have to face your stupid posts, it's really getting on my nerves now...

Thank you. It goes both ways(Xbox 360 fans PS3 fans) on this forum and it makes it very hard to tell who is telling you the truth or spreading misinformation for whatever reason. When people seem to have an agenda it makes it hard to take anything they say seriously.

 

[GB123]

To be fair, the Starbreeze guy was referring to the screenshots. Frame-rates may differ wildly (or not).

That sounds to me that your looking into something which isn't there and wasn't said.

or as multiplatform developers, they choose to be diplomatic about their words.

And you maybe right, but im sure out of all the multi-platform titles that are in development at least one developer would be on record as saying the PS3 whips the 360 ten to the dozen.

Another thing is with games like Gears of War, Too human and Mass efect looking just as good as anything i have seen on the PS3 if Edg'es statements were so true i would be seeing a bigger difference.

You could argue that the PS3 is still experienceing a learning curve but then the same could be said for the 360.

Time and time again dev's are saying the different will be less than the difference betwen the xbox and gamecube of the last gen so why do people contuine to feed this idea that one will be significantly better. (other than to make themselve feel better about the system they support)

 

[therealskywolf]

To be fair, the Starbreeze guy was referring to the screenshots. Frame-rates may differ wildly (or not).



or as multiplatform developers, they choose to be diplomatic about their words.

Well this guy wasn't really diplomatic was it? They spoked in the past Xbox/PS2, and they have spoken now PS3/360. And i'm pretty sure the guy flatout told what one sistem has better than the other and vice versa.

 

[GB123]

Some interesting info on cell..

At this year's E3 (or thereabouts) Sony proclaimed that their processor could achieve 200GFLOPS! However, according to IBM's white paper, only 155.5 GFLOPS was actually achieved (Table 4). BUT, IBM's tests used all 8 SPEs. The PS3 will only use 7 SPE's, due to manufacturing yield issues.

The efficiency of the Cell is 75.9% (Table 4), with of a theoretical peak of 201GFLOPs (Figure 5)--running 8 SPEs at 25.12GFLOPS apiece (Table 2). Similarly, the theoretical peak for the PS3's processor will be 176GFLOPS, using 7 SPEs at 25.12GFLOPS apiece. Assuming the same 75.9% effieciency, we could easily interpolate the PS3's Cell to be capable of 133.6GFLOPS.

The take home message is that with the PS3 being cabable of 133.6 GFLOPS and the Xbox 360 being capable of 115.2 GFLOPS, the PS3 is not nearly as far ahead of the Xbox 360 as we were lead to believe. we should expect relatively similar power coming from both consoles, processor power, and ease of programming all considered.

Not to mention that one of the SPE's in the PS3 are reserved for the OS and the bottlenecking of the data transfer between the SPE's and the on board memory. I see the 360 hand in hand with a gaming Revolution taking home this next round at least, if not the whole cake over time.


The person that came up with the numbers did so using information directly from IBM

http://www-128.ibm.com/developerworks/power/library/pa-cellperf/?ca=drs-#table4

 

[predicate]

That sounds to me that your looking into something which isn't there and wasn't said.



And you maybe right, but im sure out of all the multi-platform titles that are in development at least one developer would be on record as saying the PS3 whips the 360 ten to the dozen.

Another thing is with games like Gears of War, Too human and Mass efect looking just as good as anything i have seen on the PS3 if Edg'es statements were so true i would be seeing a bigger difference.

You could argue that the PS3 is still experienceing a learning curve but then the same could be said for the 360.

Time and time again dev's are saying the different will be less than the difference betwen the xbox and gamecube of the last gen so why do people contuine to feed this idea that one will be significantly better. (other than to make themselve feel better about the system they support)

The problem with this rampant speculation is that it's claiming things that just aren't happening.

Cell is supposed to be some rampant CPU monster that is so powerful it's a supercomputer etc. etc. but no games thus far suggest this is the case.

Xenos is supposed to be an incredible GPU with the theoretical shader power of an X1900XT but with double the realworld efficiency and free 4xAA at 720p but no games thus far suggest it is the case.

I suspect in a few months time when the dust settles people will realise that these consoles can't actually perform miracles

[edit]GB123: whoever wrote that compared Cell's benchmarked results vs. Xenon's theoretical max. Not a fair comparison at all.

 

[Acert93]

Some interesting info on cell..

Interesting info on Cell? How about uninteresting propaganda. Can you freaking Microsoft and Sony fanbois just STOP. The forums are bad enough with the rampant platform advocacy among the normal, we don't need this sort of dribble making it even worse.

As for your post GB123, Xenon does not have a peak of 115GFLOPs. It is more in the 70-90GFLOPs range. MS miscounted by calculating every execution path instead of "real world" peak (i.e. there are parts of Xenon that cannot co-execute at the same time meaning it is either-or and you cannot add both together). This is why the peak for CELL is in the 208GFLOPs range and not 218 as well.

Further, what the heck is up with comparing peak performance of Xenon with real-world demo peaks from CELL? Apples-to-pink-freaking-Elephants. And as mentioned earlier there is a lot more to CPU performance than just flops, and there is always the impact of architecture. This utter obsession with flops is out of hand.

I know E3 is around the corner and everyone is on their platform specific blow horn, but this is just absurd. The great contributions by ERP, Mintmaster, nAo, etc in the last couple weeks have been completely flooded with a sea of misinformation and rampant fanboism.

 

[GB123]

Xenos is supposed to be an incredible GPU with the theoretical shader power of an X1900XT but with double the realworld efficiency and free 4xAA at 720p but no games thus far suggest it is the case.

I suspect in a few months time when the dust settles people will realise that these consoles can't actually perform miracles

I guess it really depends what you are expecting from these consoles.

Personally this screen of Splinter Cell makes me believe otherwise http://www.1up.com/do/slideshow?pager.offset=1&p=&g=&tr=&mt=0&cId=3140443

 

[predicate]

I guess it really depends what you are expecting from these consoles.

Personally this screen of Splinter Cell makes me believe otherwise http://www.1up.com/do/slideshow?pager.offset=1&p=&g=&tr=&mt=0&cId=3140443

Splinter Cell has always been a 'I swear that shouldn't be possible on that hardware' game. The new one looks expectedly awesome.

 

[rounin]

Post #127 is indication that this thread has hit rock bottom. Somebody lock this please.

 

[SugarCoat]

Can you please provide links to the game code benchmarks you are referencing to be able to make this comparison? Thanks.

CELL is quite bad at integer and branch prediction, no secret and almost completely ignored because people think that the code will be so well optimized that this negative aspect will rarely arrise. What i see happening is more and more of the SPEs getting dedicated to their own specific tasks to offset this problem let alone what ever is allocated to run background programs such as the OS. I dont need to supply benchmarks to state the fact that if CELL were simply placed into the computer world as it is today it would be a total flop. Pun intended. This should be a given due to CELLs architecture. As far as hype goes apparently people totally forgot all the hype around EE vs PC that happened back in 99 and 00. CELL is a niche product for the PS3 and servers. Although i really dont think the server glory will last with the likes of POWER 6 floating around. Now speaking of the EE i'm going to ask some rhetorical questions; do we remember the quoted GFlop numbers of the EE? How about all those charts comparing it to Pentium IIs and Pentium IIIs showing its theoretical performance was far above (+3x) both? What did the Xbox contain? Did you notice a vastly superior performance experiance in either console? And this time is different how?

Really beating around the bush here. Once again, CELL is no advantage over the Xbox360 in terms of overall experiance.

 

[scooby_dooby]

Seriously...does anyone really care anymore? The consoles are in spitting distance of eachother, can't we just drop it for a while?

 

[GB123]

Interesting info on Cell? How about uninteresting propaganda. Can you freaking Microsoft and Sony fanbois just STOP. The forums are bad enough with the rampant platform advocacy among the normal, we don't need this sort of dribble making it even worse.

As for your post GB123, Xenon does not have a peak of 115GFLOPs. It is more in the 70-90GFLOPs range. MS miscounted by calculating every execution path instead of "real world" peak (i.e. there are parts of Xenon that cannot co-execute at the same time meaning it is either-or and you cannot add both together). This is why the peak for CELL is in the 208GFLOPs range and not 218 as well.

Further, what the heck is up with comparing peak performance of Xenon with real-world demo peaks from CELL? Apples-to-pink-freaking-Elephants. And as mentioned earlier there is a lot more to CPU performance than just flops, and there is always the impact of architecture. This utter obsession with flops is out of hand.

I know E3 is around the corner and everyone is on their platform specific blow horn, but this is just absurd. The great contributions by ERP, Mintmaster, nAo, etc in the last couple weeks have been completely flooded with a sea of misinformation and rampant fanboism.

360 aside as i never said that the 360 was capable of the FP calculations which were presented in that post, my main point was that Cell for the PS3 doesn't look to have as much power as everyone in this thread is assuming.

 

[nonamer]

360 aside as i never said that the 360 was capable of the FP calculations which were presented in that post, my main point was that Cell for the PS3 doesn't look to have as much power as everyone in this thread is assuming.

But the number is meaningless without context. Real world performance on both Cell and Xenon will be lower than peak performance. If Cell gets only 70% real world performance on a certain task, it is reasonable to assume similar utilization rates on the Xenon. Hence, you should not expect relative performance near what you described. More likely, the ratio in performance between Cell and Xenon will not change much going from theoretical to real world performance.

 

[AlNom]

That sounds to me that your looking into something which isn't there and wasn't said.

Excuse me? If they use the same assets, the only difference WILL be the framerate (even then, there might not be). The fact is that he was speaking only of a screenshot.

Well this guy wasn't really diplomatic was it? They spoked in the past Xbox/PS2, and they have spoken now PS3/360. And i'm pretty sure the guy flatout told what one sistem has better than the other and vice versa.

Uh... context.... he was talking about "many developers". Thanks for reading.

 

[dukmahsik]

But the number is meaningless without context. Real world performance on both Cell and Xenon will be lower than peak performance. If Cell gets only 70% real world performance on a certain task, it is reasonable to assume similar utilization rates on the Xenon. Hence, you should not expect relative performance near what you described. More likely, the ratio in performance between Cell and Xenon will not change much going from theoretical to real world performance.

of course that's assuming that cell is as easy to program for as xenon, which the dev said it isn't.

 

[GB123]

Excuse me? If they use the same assets, the only difference WILL be the framerate (even then, there might not be). The fact is that he was speaking only of a screenshot.

I'm not disputing that, what i'm saying is the developer didn't say anything about framerate so there is no point in speculating otherwise.

 

[nonamer]

of course that's assuming that cell is as easy to program for as xenon, which the dev said it isn't.

The benchmark in question is Linpack, a very straightforward test that has little to do with ease of development.