Official PS3 vs. Xbox 360 vs. N. Revolution

Wow, an incredibly dumb post, and an incredibly over-reactive response to it.
I think that was a little bit harsh wasn't it, cmmsnsense?
Though users such as morda# are extremely irritating, try and keep some of your murderous anger to yourself!

Forming a response to morda, I think reiterating Micro$oft's doctored facts is a little inappropriate for a forum with some very highly intellectual beings. At least try and come up with your own argument, even if we don't believe you.

Handy tip: Don't start posts with "I proved you wrong" because you haven't yet. If you must post such a comment, put it at the bottom after the 'proving' has neen done.

 

what a noob(morda), the kid did nothing but copy and paste, what a genoius. and like cmmnsence said, why would microsoft say there system is weaker, of cource its going to say its stronger.

 

Me too. Do you hear the storm coming? I hear they named this storm the Moderator's Storm.

Sorry, my bad. =( Sometimes people forget to flush and it clogs my brain power.

 

It was a very strong reaction now that I think about it, but it was very appropriate in the sense, that this isn't a kiddie form. You can't start crap with "I proved you wrong" and then plagiarize propaganda that both sides have. If your going to throw down statements like that you damn well better be able to back them up in your own words from third party sources. Just so were clear on memory bandwidth:

The xbox 360's main memory bandwidth is 22.4 GB/s, this does not exceed PS3's 25.6 (Sony, but realworldtech.com had it at 25.2) GB/s.

The xbox 360's FBS bandwidth is 21.6 GB/s, this, again, does not even come close to the PS3's 76.8 GB/s through a somewhat revolutionary FlexIO.

Xbox get's it's "superior" overall system bandwidth by calculating in the EDRAM, which is 256 GB/s. This provides great misconception for those who do not know any better and can only plagiarize because they don't know what an FBS even is. The 10MB EDRAM will provide insignificant in the overall performance, aside from a (damn good) graphics buffer.

I have no idea where retard got that PS3 has an over all system memory bandwith of 48/GS, because that's not even friggin close, that's only a little more than the outbound BW on the FlexIO alone.

Wvengence, Your right the 256 GFLOPS is at 4GHz, the Sony E3 press conference pegs the PS3's cell w/3.2 GHz at 218 GLOPS. Still alot more than the 360 CPU's 115 GLOPS.

 

CmmnSense:
"the eight SPEs isn't just a core, their full vector CPUs, better than cores when it come to supporting extensive GPUs and graphics capabilities" Actually no. Since each PPE core has 2 Vector processing threads vs a single thread SPE and are not even close to a PPE on a one for one comparison.

"except there shit is back by many outside sources" Actually, except for SONY and Nvidia, who exactly has confirmed any of the specifications on the PS3? No one? I thought so.

"aside from a insignificant 10MB buffer" I would be careful of those words because it is hardly insignificant and I will glady throw some basic numbers to show you how much performance it can offer.

"And at least we used sources other than xbox.com and ign.com and sony.com hey let's get away from the goddamn propaganda site and get to some actual facts. " Again, other than Sony propaganda what do YOU have to back up your claims on the PS3? Please remember that if you TRY to use the PC version of the Cell to justify the PS3 version, you WILL be admitting that the R500 will also be more powerful than the RSX (based upon the same PC tech that you use for the Cell).

"does not even come close to the PS3's 76.8 GB/s through a somewhat revolutionary FlexIO." I remember that from the IBM specs, but I'm pretty certain that SONY'S claim for bandwidth was 25.6GB. Didn't you just criticize someone for bad claims?

Since we have some idea now of the conditions that the 360 produces it's numbers why don't you tell me HOW sony got ANY of it's numbers? Since you want to compare performance then lets compare like for like. Since the R500 produces 1TFLOP ACTUAL (with 95% utilization, hierarcal Z buffer and 4xMSAA) how does the RSX fare under these SAME conditions.

Cmmnsense, I find your hypocrisy astounding. You verbal SLAM some kid for stating his belief based upon information provided by a company. That is all that you have done for anything you have spewed out. You try and dismiss anything that is negative about your precious PS3 while automatically taking as fact everything that might show it in a favorable light and deny that right to other people.

Until you can EXPLAIN to me how Sony came up with those numbers and the conditions they used, I don't want to hear a single thing about the PS3 from you. Unless you EXPLAIN how the 1.8TFLOPS of the RSX is achieved (i.e. 278GB of system bandwidth) then don't even bother to spew Sony propaganda at me.


Solargame: "why would microsoft say there system is weaker, of cource its going to say its stronger." Why would Sony say there system is weaker, of course its going to say its stronger.

Hypocrisy: Insincerity by virtue of pretending to have qualities or beliefs that you do not really have

 

Wvengence, I am going to take your post one paragraph at a time, 1. and onward.

1. I was comparing the SPEs to the xbox's PPE. Not trying to say a SPE is superior to a core. Rather the collection of SPEs are more powerful combined with the PPE. In essence the SPEs are vector CPUs, I was not trying to say they were better than a single core processor. Sorry for not elaborating for you. To my knowlege SPEs have two threads, look at your own sources (third party, not "spew" from sony sites), "The SPE is an in-order processor that can issue two instructions per cycle to seven execution units in two different pipelines." Please do not confuse the seven execution units as the seven operating SPEs, this is an SPE, not a PPE.

2. I have avoided using spec sheets from their own sources. Rather I have used the resources you provided me through realworldtech.com and arstechnia. These are third party that observed and documented the Cell conference. They also provide their sources at the end of the articles as specific members of the cell project team they spoke with.

3. It is misleading in telling people the overall system bandwidth is five time that because of 10MB of EDRAM. 256 GB/s still has to wait for the outbound of a slower memory. Foolish to think in such terms as the spec sheet is listing.

4. Check my response on paragraph two as sources go. I have not dare touch on the GPUs because there is not enough info out on the actual GPUs. I corrected my mistake in saying the PS3 would have the 256 GFLOPS rating at 4GHz, rather will run at 3.2GHz providing 218 GFLOPS already.

5. 25.6 is Sony's claim to main memory bandwidth (XDR), not it's I/O (76.8), sorry, not a bad claim yet.

6. I have yet to touch on the GPUs and Sony's claim to a 1.8 TFLOPS GPU. Don't throw that at me, I have'nt claimed/stated anything aside from the fact that no one has enough tech specs (any source) on either GPU to compare. Again, you can't use current GPUs as benchmarks, they won't even come close on the inside, just like the cell isn't like anything you can slap on your PC right now.

7. This part of your post is redundant. Go back to #2, I use outside sources, that's why I haven't touched on GPUs yet, because there is no good third party proof of either CUSTOM GPUs. I have "spewed" none such propanganda and not stated it came from it's own company.

8. Redundancy again. I HAVE NOT TOUCHED ON GPUs. Sony can claim what they want, I partially trust both sides claims, but do not back them with my own word. In the end if Sony lives up to it, they obviously have some bitchin tech with nVidia that is unprecedented and unforseen. That fact is, we don't know. And we shouldn't use current PC technology for comparison.

9. Nobody said Sony hasn't, I am not sure but I think I did say they said they're the best too, if not on this forum site, another. Both sides can do it, but the cell has alot of spec confirmation from alot of companies and has released much of it's specs at the IBM conference and has displayed it. Xbox's CPU is unproven to the public aside from playing games on MTV.

10. I do not appreciate your accusations, which are indefinitely false. You are trying to put words in my own mouth and tell me that I am getting my facts from Sony, when I am not.

 

On #5, it's not just a claim. They were presented at the IBM conference, and is not just Sony's sayings, but the entire cell team and the companies behind them including Sony.

 

Being fair and just, with no bias at all, I believe that cmmnsense's argument is holding up better at the present. I also agree with him... Let's just hope he doesn't get banned for slamming the noob earlier!

To toiletman: apology accepted.

 

FSB=Front Side Bus... Connention between the CPU and northbridge controller (in this case, system RAM)

System Bandwidth
Main RAM 25.6GB/s
VRAM 22.4GB/s
RSX 20GB/s (write) + 15GB/s (read)
SB< 2.5GB/s (write) + 2.5GB/s (read)
from ps3.ign.com based upon Sony's announcment.
Main RAM 25.6GB/s Seems to be about 51.2GB/s short.

While we are at it, let's look at your 76.8 number "The 12 byte lanes thus provide 44.8 GB/s of raw outbound bandwidth and 32 GB/s of raw inbound bandwidth for total I/O bandwidth of 76.8 GB/s. Furthermore, the byte lanes are arranged into two groups of ports: one group of ports are dedicated to non-coherent off-chip traffic, while the other group of ports are usable for coherent off-chip traffic. It seems clear that Sony itself is unlikely to make use of a coherent, multiple CELL processor configuration for Playstation 3"

Since you know what it is your quoting here, let me fill in everyone else. The Coherent/Non-Coherent Off Chip traffic it talks about? The ONLY thing that is used for is CPU coherency and Dependant Thread calculation. To put it in more friendly english, this 'somewhat revolutionary FlexI/O' bandwidth is used ONLY to talk to another Cell processor in the same system. If there is no other cell, there is no use for that bandwidth, PERIOD. So while yes, it 'technically' has this, the bandwidth mentioned goes absolutely NOWHERE. I think that qualifies as a bad claim.

"The SPE is an in-order processor that can issue two instructions per cycle to seven execution units in two different pipelines" If you read your own quote here, it is saying that the SEVEN SPE's can only execute 2 instructions per cycle. Thereby losing the concession they could take 7 at a time.

"It is misleading in telling people the overall system bandwidth is five time that because of 10MB of EDRAM. 256 GB/s still has to wait for the outbound of a slower memory. Foolish to think in such terms as the spec sheet is listing" Sony no doubt used simular logic for every number they pulled. Until you explain to me HOW those number are generated, they are just as foolish.

"That fact is, we don't know. And we shouldn't use current PC technology for comparison." This is YOUR quote from your reply. The sources from Arstechnica & (Especially) Realworldtech as based upon IBM's 4Ghz PC version of the cell. First, you tell me we can't use PC tech for comparison (because the RSX GPU would be shown as easily inferior) but then go on to use the PC Cell as the basis for claims of the PS3 Cell.

"I do not appreciate your accusations, which are indefinitely false." Funny, if you say I can't use PC tech as a comparison, then your ONLY comparison not from Sony is using PC tech, then your right back to Sony as your only source. I'm pretty sure that falls under the 'beliefs that you do not really have' portion.

 

An impressive bicker, but I think my mind is made up at the moment, slightly in favour of Sony.
It's possible to gleam this much:

PS3: More processing power
360: More graphical power (despite 10MB of RAM)

 

My thoughts too, Sammorris. Although the 10MB does more than you think.

 

As far as you SPE claim, it's seriously misconstrued.

Figure 4 - SPE Organization
"The SPE is an in-order processor that can issue two instructions per cycle to seven execution units in two different pipelines. Typically, each instruction makes use of 3 source operands to produce 1 result."- This is a single SPE, it has two threads, it's own LS, and seven execution units. Just because seven is the same number of operating SPEs in the cell doesn't mean that this magically just became the PPE, besides, the cell version were looking at from IBM has all 8 operating. Damn, why are you so convinced even though it says SPE and the SPE has 7 execution units, it takes in two instructions, and 3 source operands are necessary for each result. Each of the seven SPEs have seven source operands. The SPEs are fully independent of the L2 cache and the PPE.
http://www.realworldtech.com/page.cfm?ArticleID=RWT021005084318&p=5

Read that and look at the pretty pictures. The cell processor operates so XDR>MIC>L2>PPE and simultaneuously XDR>MIC>SPE
The SPE takes in two instructions at once from the system memory w/out the PPE. Now, instrution can also go as follows, XDR>L2>PPE>EIB>SPE (shortend up by skipping XIO etc., less typing) and SPE>EIB>SPE... Eventually all 18(maximum if MIC sends two instructions to each SPE and PPE, and they are processed and out EIB) processes through many different paths (the PPE>SPE and SPE>SPE, aren't frequently necessary paths, rather MIC>SPE>) get out EIB>BEI>FlexIO>redwood. INSTRUCTION THAT COMES INTO THE CELL FROM A VIDEO GAME COMES IN THROUGH XIO AND EVERYTHING THAT LEAVES GOES OUT THROUGH FLEXIO. FlexIO is not to connect more Cell processors. If you do connect multiple cells, I/O does control communication of the cells to system managment storage and networks(in a multiple cell design). The design IBM revealed showed the cells connected "gluelessly" to each other, not even through any I/O. Sorry only slightly a bad claim, just a different bandwidth. 76.8 GB/s is the processors outboud to redwood then we need to know what redwood is so we can see the bandwidth getting to the GPU.
http://www.realworldtech.com/page.cfm?ArticleID=RWT021005084318
http://www.realworldtech.com/page.cfm?ArticleID=RWT022805234129

"is used ONLY to talk to another Cell processor in the same system."- Now there's a bad claim.

"Realworldtech as based upon IBM's 4Ghz PC version of the cell"- It's the same aside from the Voltage going through it, and all 8 SPEs running, I already said the PS3 will have 218 GFLOPS, not this cells 256GFLOPS.

"it is saying that the SEVEN SPE's can only execute 2 instructions per cycle. Thereby losing the concession they could take 7 at a time." - Bad claim, it doesn't hold against actual analyzation of the acticle, and purely to add to that, it would make no sense to have seven operating SPEs and only issue two of them at a time. What a waste of money and research, common sense (Let's leave that one alone, It's the best word for my point) would tell you they wouldn't do this, and it's confirmed that it's not the configuration.

 

Cmmnsense:
Save your breath (figuratively). The Cell information you talk about is PC technology and therefor has no relevance on the Console discussion.

 

Overall, it doesn't really affect anything, but I am making my point. If you want we can just stop, and just say the Cell will be slightly superior, but both systems will be close? It's up to you.

 

It is the same cell though, just difeerent clock speed, and the PS3 has one SPE reserved.

 

For the PC Cell.
http://www-1.ibm.com/industries/telecom/doc/content/bin/tc_isscc_10.2_cell_design.pdf
Looks like we meet in the middle. IBM lists it as 1 PPE & 8 SPE's capable of supporting 10 simutanious threads. So, no, the SPE's aren't dual threaded. However, I can't find anything on the Cells regarding the thread input so I'll settle for the fact that they accept threads independently.

Otherwise, it doesn't matter to me what you want to do. I'm just saying that either you allow the comparison to PC technology for everything or you don't for everything.

If you choose to not use PC tech, then we can't really debate the specs because we have already confirmed that both sides are biased. That leaves services and online (which I must say, the 360 has an advantage).

 

"we could just stop" - we're making progress!!! lol

cmmnsense I agree with your claims that the cell will be stronger as I have said on numerous occasions, but for someone who slams noobs in unhealthy portions, you could give yourself a bit more credibility by not double posting and by proofreading. You'd be surprised what a difference a clean post can make to your apparent intelligence.

 

Double posting? You mean puting one right after another? There's nothing wrong wit that if you forgot something. I can't remeber where, and I wish I could support it as much as you would want vengence, but I think it was the E3 Sony conference where Kurtaragi or however you spell it said that the cell in the PS3 is the one and only, but they just have it running at a different speed and reserve an SPE, I already said thi a few poss above. Also, in all reality it isn't PC technology, 1. because you can't buy it yet (PC- Personal Computer) 2. we may never beable to buy for a PC depending on how it does as a worksation and the PS3. I don't consider it comparing current PC tech at all. I end up reiterating things alot because the posts after my posts (morda8) don't seems to have read exactly what I said.

We do meet half-way, nice IBM link, that actually does alot of clarifying where the others left stuff in the open for assumptions.

I agree that the xbox's GPU looks better, but I want to wait to discuss those until ATI and nVidia gives us more on the specific GPUs.

As far as features go, we don't know what PS3 is doing for online this time, but live is pretty hard to beat. Features, I feel more strongly about the PS3 features, but features comes down to what each person wants. If microsoft will publish windows for PS3 (linux is apparently ready) then the PS3 will definitely have most of the feature that xbox will streaming media from your computer. Sony I hope can do that too. Features are more to each his own, and online for me personally is purely dependent on the fact that PS3 will have socom 4 (score).

Sammorris, I don't proof read these, it's just a forum site afterall.

I think it'd be best to just post news up here and then when new press releases give us more then we can discuss further. I think we've covered everything pretty well.

If anyone here didn't know, the PS3 is going to use procedural sysnthesis too. They just can't use it on 3-D character skinning, because MS patented that, damn.

 

xbox is going to rule period...

 

Anger coming out... lord grant me the serenity not to find and mame davidl1l...