Xbox 360 vs. Playstation 3 specs comparison

[Reanimated]

Microsoft XBox team member Major Nelson (http://www.majornelson.com/2005/05/20/xbox-360-vs-ps3-part-1-of-4/) has a comparison up of the 2 consoles with pretty charts and fancy numbers. Xbox 360 comes out ahead in 3 of 4 comparisons, including shader performance.


Let the sceince fair wars begin!

[chechenepiphany]

wow, the memory bandwidth diff. really surprised me. i thought they were neck and neck on most things. or at least not that far apart.

[Borys]

You did mention that he works in Microsoft and on Xbox 360, didn't you?

Talk about being unbiased.

[Swick]

You did mention that he works in Microsoft and on Xbox 360, didn't you?

Talk about being unbiased.
First sentence of the article. Kinda hard to miss.

[Vjornaxx]

wow, the memory bandwidth diff. really surprised me. i thought they were neck and neck on most things. or at least not that far apart.

MS has really been talking up their "no bottleneck" design philosophy. What good is a ridiculous CPU or GPU if it can't get the data out fast enough to the other components?

[happycat]

This comparison of which console has the longest hardware pensi is going to be one of the last things on people's minds when they actually do come out, and basing your purchasing choice on which console can squirt out prettier graphics is downright stupid. Features and games, thats what counts. The graphics on the PS2 look like ass, but people still buy them, why? because someone may want to play GT4, GTA or the newest FF, sure the game looks like crap but that doesn't stop you from enjoying the story in FF, blowing junk up in GTA or navigating a track in GT4.

In my case I love racing games, one of the few genres that are best played on a console. The xbox has the best selection of them that are also multiplayer enabled, hence why I bought it a year ago.

[Lexicon]

I'd be a lot more likely to believe those charts if they gave any kind of factual backing on how they come to those numbers other than, we say it is like this.

[EvilBob46]

Wait a minute. Apparently, this guy works for Microsoft and is doing "a real world performance comparison" without actually having access to the PS3 hardware and is using estimated numbers from the spec sheets, making nice bars and graphs out of them in an attempt to make them look like actual benchmarks.

...

You've got to be kidding me. And the funny thing is: the average reader is going to look at this and think it's actually a meaningful comparison and go "gee, the PS3 is actually less powerful." Is EvilAvatar in any way funded by Microsoft? Be honest now.

[Borys]

I know you are into that kind of tech stuff, Reanimated, but it's one thing when a group of independent analysts write a report on something and another when a guy that works for one of analyzed companies does it.

He won't be 100% objective, he will always try to sell you his product and show it in a better light, ahead of the competitor.

Racist jokes won't hide the fact that this is a wrong edited newspost. It shoud say WHO this Major Nelson guy is - an analyst, writer, web editor, Xbox designer I don't know.

Pop on my ignore list you go, little blind racist.

[Ernst_Jager]

yeah but all of PS3 demos were 'up to specs' so what is the difference really? How can anything be according to specs when neither system is even finished in design and hasn't been released yet?

[kickmybum]

basing your purchasing choice on which console can squirt out prettier graphics is downright stupid.

I'll probably buy whichever console that can squirt out prettier graphics.

[Reanimated]

Polish is a race?

Give it a rest oh "Captain of the Xbox Death Squad", I've seen you post crap with FAR less factual information.

[dr_qwandry]

and i'll buy one that has games that are good instead of just looking good

[Rommel]

I know so very little about the subject that alll I will say is that I love the promised features for the XBox 360. Also, I think that whether or not the cell works perfectly now - it will be the model of computing in the future. Hive mind machines dividing up tasks... waitaminute. Oh my God! I forgot the lessons of Hollywood!

[ColdForged]

I believe Reanimated probably assumed people knew Major Nelson by name. Certainly far from the least biased perspective, but if you can find fault with the math, please contradict.

I'm interested in the EDRAM bandwidth and the implied "no penalty" operations it provides. In my mind from what's available in terms of information today, Microsoft really does seem to have the most balanced system for gaming. Until someone can tell me how developers are going to put those 7 SPEs to work and keep them churning, I'll take the 3 general purpose cores of the 360's PPC.

In the interest of disclosure I own an Xbox, plan to own an Xbox 360 and might buy a PS3 depending on how Motor Storm, Killzone 2 and perhaps other exclusives turn out.

[bapenguin]

haha...that memory bandwidth chart is hillarious....MS is just playing the numbers game that Sony played....

Here's how it breaks down...from a reliable source:
The Cell Processor is faster in FP, slower in Integer (shown on Major nelson's site).
The problem with Cell...and there's no way to prove this right now, it's all in theory....it's in the way it passes off information. As the main processor passes off processes to it's cells, it needs to wait for them to finish. The ineffeciency comes when you end up when there's the delay in the MAIN processor waiting on the cells. It's sort of like memory timings on a PC. Simplified form: CAS 2 ram is 2 clock cycles per access of memory while CAS 3 is 3 clock cycle per access. So theoretically...CAS 3 is slower...but you don't SEE this until you approach the limits of the memory...then it bottlenecks. Same thing with the Cell....ok I'm just rambling...just give us the friggin games.

[Doomsday]

The announcement & specs of the PS3 have put the rah-rah Xbox squad here at Evil Avatar on full-on damage-control mode, haven't they? I don't think I've ever seen as many articles with the same theme posted by Reanimated and Bapenguin in a row in the history of the site.
What's the news this time? That a guy who WORKS on the Xbox 360 thinks it's more powerful than the PS3? Next you'll be telling me Ken Kutaragi thinks the PS3 is more powerful than the Xbox 360 based on a photo of the system's case he saw on a camera phone.

Bring on the "news", Reanimated.

[bapenguin]

You've got to be kidding me. And the funny thing is: the average reader is going to look at this and think it's actually a meaningful comparison and go "gee, the PS3 is actually less powerful." Is EvilAvatar in any way funded by Microsoft? Be honest now.

2 things:

1) No specific information about the benches were posted in our article, simply a link to Major Nelson's. Someone then reading the article would havee to read the part about him being an employee.

2) no, we aren't funded by them. Though Ensemble studios (part of MS) did have ad space on here for a while.

[Lexicon]

I like how he includes the 256GB/s into the total memory system bandwidth to beef up that chart.

From Anandtech:
Remember the 256GB/s bandwidth figure from earlier? It turns out that that's not how much bandwidth is between the parent and daughter die, but rather the bandwidth available to this array of 192 floating point units on the daughter die itself. Clever use of words, no?

As for the shader operations per second its hard to tell where his numbers come from since ATi says Xbox360 GPU does 96ops per cycle and Nvidia claims the RSX does 136 per cycle.

In the interest of full disclosure I own all 3 current systems, as well as all systems back to the Odyssey 2. I just think the number fudging games MS and Sony are playing are silly especially when they are this obvious.

[raVen]

Definately take this information with a grain of salt, there might be some truth to parts of the article, but I find it hard to imagine that good chucks of it aren't fresh MS propaganda.

The hardware arguments aren't going to get us anywhere, it won't be until at least the second or third generation of console games where we can start to get a clear idea of which is more powerful.

[AspectVoid]

This is all pointless and stupid. When are people going to get it through their heads that the damn specs don't matter one bit. What matters is which system has the games YOU ENJOY!

I am so sick and tired of all of these dumb, pointless speculation articles that are coming up everywhere. The fanboys are out enmass for all systems, and there is no getting away from the freaks. Mind you, if this had come from a sony rep, I'd be posting the same thing. I just want it all to go away.

[ElectricMonk]

yeah it seems a little biased and useless. adding up all the bandwidth numbers means what?

there are enough fundamental differences between the two systems that we'll never know until after games are being made.

i still think that sony has something they're not telling us about the rsx because of the lack of details released about it. if they're going to give outrageous claims and never explain them it's always going to be suspect.

[Lagrius]

Is there any way to make a poll seeing who's getting one, two, or all three systems?

[bobbler]

I don't mean to sound negative, but that comparison is truly one of the most skewed I've ever seen. In addition to just being a pathetic jab at the competition, its on a friggin' xbox developers site -- you can't expect an untainted comparison.

Many of the charts are not showing the entire picture (remember what we all learned about statistics?)

For example, the memory bandwidth one? If you can't see whats wrong with that then you've got issues, and probably shouldn't be near the writing a blog about technical aspects of a console -- you don't add aggregate bandwidth like they are beans or something; bandwidth of one section can't just be added to another, it doesn't make sense no matter how you spin it. Major Nelson should know better, this is pretty sad.

Its an attempt to skew the picture, thats all.

I say this again, this is not even close to valid.

[Draft]

Yeah but which console will increase gigaflopping by an order of magnitude paradigm shift?

[*Legion*]

There's some meat and some pomp there.

The memory bandwidth graph is a bit of pomp, although it IS a real advantage for the Xbox 360 having the super-fast EDRAM for the antialiasing calculations. I think the real point to be made, though, is that the two systems are otherwise on par with memory bandwidth.

I think what people should take away from this, though, isn't necessarily the idea that the Xbox 360 is more powerful, but that the two systems are awfully close. What's hilarious is to look at these specs and see SO much that is almost the same, and then see each side twist the 10% that they can twist and try to show themselves on top.

This isn't going to be like last time. There won't be a PS2 < Xbox power differential this time. It's going to be a pretty level playing ground.

[Kentor]

Until someone can tell me how developers are going to put those 7 SPEs to work and keep them churning, I'll take the 3 general purpose cores of the 360's PPC.
NOTE ON THE ARTICLE: I don't see where he's coming up with these graphs.... they don't seem to correlate with the marketing or specified performance numbers provided by Sony and Microsoft.

It's not nearly as hard as it sounds, but it requires a different mentality than developing for a massive instruction window preempted processor. It's important to know why they're essentially using SoC DSPs instead of just making a multicore processor first though.

PS3's Cell is structured to allow for easy RISC like DSP processing... each SPE is statically scheduled, has low instruction latency, no instruction reordering, etc. The basic idea is to throw away all the traditional blocks used to increase instruction level parallelism reducing microinstruction architectural complexity in an effort to reduce the massive instruction latency found in most modern processors.

Don't think of instruction level parallelism as a way of increasing the processing element utilization, but rather as a method of hiding processing memory instruction latency. The gap between instruction execution and external RAM operations is the single most significant cause instruction latency. As the gap between memory and processing increases, the more complex these mechanism must become; cache sizes must increase, pipelines must be deeper, control logic must become more complex, etc. All of this just increases the pipeline latency. The SPEs significantly reduces this instruction latency using an approach very similar to early RISC and modern DSPs.

Why is this important? DSPs are used because they are extremely fast with MAC operations, have exceptionally low latency, and are extremely predictable. There is no need for complex synchronization, data throttling, or dynamic scheduling, greatly decreasing software complexity resulting in very low overhead -- even in the case of multiple SoC DSPs on the same die (this is actually fairly common in the world of DSPs and embedded RISC processors). This results in quite a few DSPs which can process far fewer operations per second at peak than a general purpose processor vastly exceeding the performance of that same general purpose processor (though many modern DSPs easily exceed any general purpose processor in terms of MAC performance). Some modern DSPs operate at the same frequency as external memory resulting in no latency at all (granted they aren't clocked in the gigahertz range). Also, in terms of instructions, it is very often far easier to keep a DSP at maximum utilization than a general purpose core due to the hardware's predictable nature and overall simplicity.

Keeping the SPEs at high utilization shouldn't really be a problem for computational tasks -- especially not real vector space calculations, though it might take a bit for game developers to get used to DSPs and I think that is a significant concern.

EDIT:
-I don't believe I actually mispelled correlate

[falak]

I just want to point out that Microsoft are always pulling this shit in the OS "wars".

[EvilBob46]

1) No specific information about the benches were posted in our article, simply a link to Major Nelson's. Someone then reading the article would havee to read the part about him being an employee.

No specific information about the benchmarks was given? How about "Xbox 360 comes out ahead in 3 of 4 comparisons, including shader performance." If you're going to note the specific results of a benchmarking article, you might as well, for the sake of responsible journalism, state how these should be taken with a grain of salt. How many people read this site and only skim over newsposts or aren't tech savy enough to bother reading about CPU floating point performance and chose to instead just take your headlines or short comments for granted?

[Taco]

Did Microsoft consult with Apple on these benchmarks?

[Mephistopheles]

Did Microsoft consult with Apple on these benchmarks?

Champagne comedy!

[Hellstorm]

I did some real world benchmarks, the Atari 2600 crushes both of them.

[Syrinx]

I did some real world benchmarks, the Atari 2600 crushes both of them.

It's not backwards compatible though.

[Balthasar]

I keep thinking the threads on EA could not get more retarded; then we get posts about benchmarks done without hardware. No really, this is worth discussing. Perhaps PC users should start considering more closely what Nvidia and ATI say about their GPU's, since, you know, you can just take their numbers and perform benchmark tests without EVER HAVING THE FUCKING HARDWARE.

I think it's threads like these that keep me coming. How much lower can it go? How about we start comparing the graphics of DOA 5 and Tekken 6? Just as valid as anything else.

[Taran Wanderer]

I think the point of Major Nelson's blog entry is not to actually give accurate numbers comparing the two machines. He's simply trying to fight the lies of Sony with lies of his own. Sony did a masterful job of duping the major E3 journalists into reporting that the PS3 will definitely be at least twice as powerful as the XBox360. Sony made these claims without having any hardware to show, and topped off their propaganda extravanganza with purported "real-time" video of game footage. It's hard to blame Sony considering their success in sabotaging Sega with the "emotion engine" nonsense. Sony is trying to convince people to pass on the XBox360 and wait for the PS3. Microsoft is trying to counter this campaign. A smart consumer will simply wait and see what each console has to offer.

[mister_slim]

There seems to be a lot of arguing over whether Sony or MS lies the most (which is probably a victory for Nintendo, if you think about it). How about this: Longhorn and the PS3 both launch in '06, so when they hit we compare the misinformation.

That would be about as useful and accurate as these comparisons.

[bobbler]

After reading this again, I can't help but comment more -- this is just too funny. What is posted on major nelson's site is far more 'silly' than anything Sony has ever said, by landslides. This is the toy-story comment of this generation. Its absurd. Chuckle worthy, even.

Claiming xbox360 is xbox1.5 is an opinion, childish maybe. Showing a few videos of what games (probably will, I would expect both consoles over time to get close) might look like over time is silly marketing. However, making stuff up (not even using the specs your own company released in some cases, and just coming up with numbers out of nowhere for others) and putting it into graphs is flat out lying.

The difference it that these guys are uniquely qualified to do so, and can cut through the smoke and mirrors to see what the real deal is.

He even claims this is the real deal. The REAL DEAL folks. These graphs are cutting through the smoke and mirrors. Oh my god.

The first graph hes just using counting as a method of actually displaying real info. "we have 3 cores, they have 1." -- 1.0 and .33 then!

The second graph I don't even know where he came up with that. Maybe Gflops? 115.2 vs 218 ? Who knows. Regardless, even if it is in Sony's favor its not a very accurate way of showing much besides how you can compare two numbers with a graph on a scale of 0.0-1.0.

The third graph, and this is where it starts getting real fudgy, comes out of left field. The numbers microsoft released where 48 b shader ops / sec, not 80 b shader ops / sec. You figure it out, I have, really, no idea where this number came from (for either PS3 or Xbox2).

The fourth has to be my favorite. Adding up numbers of whatever he saw on the spec sheet apparently, eDram included. This would be akin to adding up the dram bandwidth that each SPE inside the cell has -- that'd be 7 x 128bit x 3.2ghz, I think. Those are memory too, they aren't technically cache.

All jabbing aside, this man is nuts -- of course so are a few of the PR people from Sony. This is a perfect example of marketing at its best (worst?). The waiting game begins =/

[bobbler]

It seems Microsoft is perfectly happy spreading around misinformation also (http://xbox360.ign.com/articles/617/617951p1.html)

[megatron666]

I think bobbler makes some really good points. All these charts are assumptions. There isn't any hardware out there to do a real world comparison. With that in mind, let me throw in my 2 cents.

First graph, it is saying that the 360 has more general purpose processing power. Ok, I'll buy that, but I don't care about running general purpose applications on my GAME console. I just want to play games. This is specialized if anything. So, I think this chart is all BS.

Second chart, PS3 wins. But, again, this is just an estimation. There is no real hardware from either company to back up these numbers.

Third chard, the 360 (supposedly) has more GPU processing power therefore their shading technology is better. Also, "its innovated features contribute to overall rendering performance." What features? I'm sorry, but a lot of hand waving and tales of "more MHz means more power!" doesn't convince me. Intel said the same thing.

Last chart, total system memory bandwidth. The 360 is saying it has "256 GB/s memory bandwidth to EDRAM" and this is where such a huge number is coming from. But what is this EDRAM? Acording to this link:
http://searchwinsystems.techtarget.com/sDefinition/0,,sid68_gci213927,00.html
It looks like the EDRAM is cache on a processor. This sounds like some crap to me. Sony didn't even post how fast it's processor cache is. So, lets just take it out of the equation and compare the ps3's total video and system ram to the xbox 360 total video and system ram.
ps3: 48GB/s
360: 44GB/s

This seems much more reasonable. I don't see how ram speeds could be so much faster unless the memory is directly on the chip. Then, at that point, it is going to be high speed registers or cache.

Now, I'm not trying to claim that the ps3 is more powerful than the 360. What I am claiming is that what this guy has posted is total garbage and should not even be read. Wait until the systems come out.

[bobbler]

I posted on his site, been waiting for a reply from him (although I have my doubts it'll happen, he tended to only respond to people without any sort of proof in their posts). I've already had to fend people off like crazy

I'm surprised more people aren't upset by this. Its far more disguisting than anything Sony has done PR wise. Its flat out blatant lying and claiming it to be the 'whole truth'.

This certainly is not the way to win customers, but it is funny.

There were far better ways Nelson could have gone to showing the ways Xbox360 is really incredible (pictures of AA and Non AA for example, and how those will come at almost no performance cost).

[jace024]

I think overall performance of a console is extremely important. That being said, I also think there a several other things to consider before purchasing - this isn't a PC where they're all the same platform and performance is king. Game titles are also a factor, but besides a few franchise titles, most games are released for all platforms. I think there are still other important things to base one's decision on.

The current round of consoles has a clear winner with the XBox. I'm not a MS advocate either - I had a PS1 and loved it! Of course everything is subject to opinion, but it's clear that like game titles play smoother and look better on the XBox. I don't think this is what gives it the edge, though. It has a built in 4 way multi-tap (essential w/ friends and sports games), online hardware was built in from the start (and XBox Live was done well), it offered support for HD, and probably most important was the addition of local storage for several reasons: I've never bought a memory card and have saved games from every title I've ever bought, borrowed or rented - even season stats from sports games that used to take up an entire memory card. Little extras like personal sound tracks for games and such are also neat, but the kicker is the load times. The XBox load times are hands down the fastest ever on any CD/DVD based console. PS2 load times are worse than the PS1. Of course it's loading much more data, but yikes! XBox load times absolutley SMOKE PS2 and it's painfully apparent when I play PS2 at a friend's house.

With the PS3 and XBox 360 - the performance race is much closer. (We won't know exactly how close until the actual units are released, unless you believe biased and unsubstantiated claims like the comparison at the beginning of this thread.) It seems likely that it will be the intangible extras that will boost one far enough above the other to be the clear victor.

[TheBigL]

Hehehe I Read it so long ago. :D
Guys you all understand that its all complete BS right? How could they compare something that is not out yet, yet alone READY!!! (Not to mantion done COMPLETLY BY MICROSOFT!!!).

I read it in MegaGames (http://www.megagames.com) site, I advise you to read it there as well since there is a twist in the end ;)

Read and laugh here (http://www.megagames.com/news/html/console/xbox360vsps3-themicrosoftstory.shtml) They posted it in: Monday May 23, 2005.