New Cell processor previewed at chip show - runs twice as fast.
Nice anyone buying a PS3 up to that time will get the retarded version then... Now even more ppl will wait before they buy. Nice move Sony!
The Cell has more uses than just PS3.But this is great news as the PS3's Cell will get a die shrink to 65nm.
Mart is right on. Now my two $600 PS3's are subpar? That is crap. Talk about screwing over your customers.
They'll just put a 3.2GHz cell with this 65nm die fabrication on future PS3's, that's all. Just like they do with the 360's Xenon, btw :) So are you disappointed to have a retarded 360 then ?
Small difference Raist They say the new Cell will be 'twice as fast' (if that's true, because I've heard more about the cell that isn't like that in reality), I haven't read that about the 65 nm 360 CPU that would be used in the future
Yeah i know. I'm just sayin that i highly doubt they'd use a twice faster cell in future batches of PS3's. That wouldn't make any sense. What is probably going to happen is that they'll use 3.2GHz cells with a 65nm die.
The new cell chip might be able to catch up to the processing power of the 360's CPU.
so this new version is doing what, over 400 Gflops? Scrumptious over here obviously can't read numbers since the 360's CPU only pulls off 75 Gflops. A paltry number, especially with Quad-core processors already kicking its ass. Like we said before, Emotion Engine was just the starting point, but Cell's the step in the right direction. The next-gen Cell chip will make it into Playstation 4, and was previewed to have 4 PPEs and 32 SPUs. Sony also has a new development contract with nVidia, so that will probably have great results yet again.
These chips might be implemented in the new slimmed down PS3s that I'm sure will come out like the PS2 and PS1 versions did. More likely though, this is just the next step towards PS3. With IBM and Toshiba working on Chip, nVidia working on the GPU, and better blue diodes coming, it'll be a good machine. Its obvious that both Sony and MS are already thinking next-next gen and the two machines they release will be phenomanel (spelled wrong).
The GFLOP numbers for the PS3 were calculated based on 8 running SPE, so the fact that the PS3 uses only 6 SPE for game applications lowers the peak theoretical even further, as majority of the floating point work on the PS3’s CPU is done via the SPE. Each SPE has a peak theoretical of 25.6GFLOPS. So the total peak theoretical performance for all 6 SPE would be 153.6GFLOPS, but why is that number also not achievable? In IBM’s controlled testing environment, their optimized code on 8 SPE only yielded a performance number of 155.5GFLOPS. If it took 8 SPE to achieve that, no way 6 will be able to and that testing was done in a fashion that didn’t model all the complexities of DMA and the memory system. Using a 1Kx1K matrix and 8 SPE they were able to achieve 73.4GFLOPS, but the PS3 uses 6 SPE for games and these tests were done in controlled environments. So going on this information, even 73.4GFLOPS is seemingly out of reach The asymetric design of the chip and bottlenecks is why the CELL will never reach the actual game performance of the 360 CPU. Then throw in the fact it has to help out the weak GPU and shuffle data back an forth through the bus for 2 different types of memory... The CELL is a great chip for TV's and cable boxes. Not great for gaming.
about a new cell processor being put in a PS3v2. all i care is if and or when they are going to make Cell processor compatable with destop computers in the near future. i mean, [email protected], thinkin of how many plugins i could for my DAW Setup is almost infinite(or more than i would ever use even on my must complex of sessions/projects) 4 PPE's and 32 SPEs all runing at 6.GHz. that insane. the most powerful DSP i've seen ever of hear of has been clocked at 650Mhz
There you go http://www.mc.com/cell/prod... have fun :D
kinda cheating? i know this is 'war', but ive never heard of a company doing that. its like coming up with a new consloe alltogether... it also seems a bit desperate in my opinion and makes it feel like the ps3 isnt the console they wanted and also makes it feel that it was rushed. and The Mart is right, if they do that then the early ps3 buyers would be left out and would have to shell out another $600 (or more).
thanks man, i probably never would have found that on my own. BUT...... THAT THING IS HUGE. the only case that the supports is the one from hp. Im not a HP man
Do you even know what a Matrice is? You selectively picked a benchmark that videogames don't apply to in any way, shape, or form. Congratulations for showing your ignorance. The same benchmark page shows that Cell, when it comes to graphics, is 30x more powerful than G5 cores with VMX (3 of which are in the 360), so your argument is null. Microsoft doesn't want to publicly benchmark the Xenon because the evidence shows that Cell already creams it. We have tons of video evidence that Cell kicks ass at everything from graphics rendering, to physics, AI, and sound processing. A lot of haters still claim that PS3 only has 4~6 SPUs to rely upon and is thus its downfall; the same people also claimed that the SPUs aren't real processors, so they're showing a bit of contradition in their words. Hell, they can't even get the supposed number of working SPUs consistent. In reality, they have no clue what they're talking about because unless Sony publicly states something about how the PS3 OS works, we know nothing. Rumors and 'anonymous leaks' don't prove anything. ----------------------------- -------- To answer a general question about the new Cell processor, companies update their processors constantly. Sony, Toshiba, and IBM know that they can't wait 5 years to release a new chip and instead want to keep pushing Cell forward. This new version of Cell will NOT be put in the PS3, so don't worry about being left behind. They're simply preparing for the future and attempting to make their way into the PC market. The reason the original Cell CPU was such a monumental task is because it gets rid of most of the bottlenecks and issues that Intel and AMD's x86 processors suffer from. Even the newest DuoCore Pentiums and FX 64s are based on designs from 15 years ago. It's going to take a lot of hard work and cooperation from the entire industry if STI wishes to put Cell into everyday PCs due to backward compatibility issues. Windows would have to be rewritten from scratch.
has one PPE and 8 SPE's (7 off which are running). each SPE has its own local store of upto 4MB. (but the one in the PS3 is 256 KB, so low that you can call legally retarded). the cell can be ran in a multitude of ways. the PPE can dish out orders to the SPE and and monitor their progress. the Cell can be a Hydra (9 independent heads). and to other i cant remeber right now.
but once they upped it to 256 they ended up getting something like 4 or 5 times the original performance. 256 KB of on-chip memory per hardware thread is actually pretty good. The SPUs don't use traditional L2 cache so they know exactly what they have in memory at any given point; the speed of cache with the benefits of RAM. It's better than the Xenon CPU's situation of 133 KB of on-chip memory per thread, but hopefully they can up the ante to something like 1 MB per SPU. Now THAT would be insane. It's a space and power consumption limitation for the 90nm version of Cell, but from the looks of it that issue is being dealt with fairly quickly. To Djbassay: The other methods that Cell can do is the streaming model where each SPU works on data and sends it over to the next SPU for a different function, and then to another SPU, etc. ---------------------------- If Sony puts in This specific version of Cell into PS3s, it wouldn't be the first time they've upgraded their console CPU. PS2's EE chip was originally 297 Mhz, but was quickly upped to 333 Mhz.
May I ask why we should take your version of how the PS3 Cell works and how powerful it is? All I've heard is you quoting numbers from Sony's PowerPoint presentation and running with that. Every company exaggerates their numbers by using benchmarks that gives the greatest results. Do you have official documentation/white paper on the PS3 Cell? Just curious as to what makes you the expert. I would also like to know what videos you are referring to that proves the Cell "kicks ass" at everything. Or are you just referring to videos of games?
yeah DJ might just be referring to videos of games as proof that the cell "kicks ass", when actually so far from everything we've seen the Cell hasn't proved its done anything to make GAMES LOOK BETTER than what the 360 offers....it's funny(he's delusional) because at the moment most 360 games look better, so all the numbers in the world can't disprove what "normal" people/reviewers can see with their own eyes. Now I'm not bashing the Cell, because it is an awesome chip and I don't believe that the PS3's slightly lower graphics so far has anything to do with the Cell "underperforming", but also don't think it will ever make PS3's graphics significantly better than the 360's because of its "raw" power....because I believe it always comes back to "WHAT THE DEV'S DO WITH IT", -which in fact is a combination of many different variables (CPU, GPU, SPU's vs Cores, etc.) but most important is which format can be OPTIMIZED the BEST for GAMES, -in other words this kind of reminds me of the specialized power units some of the tech guys work on at my job....there are 2 versions, the MP-3 & MP-5, the specs for the MP-5 are definitely higher than the MP-3, but the tech guys love the MP-3 because of its ease of use and overall efficieny at delivering on average more ppu's than the MP-5's. Yes the manual says the MP-5's are capable of 821 ppu's but under real world conditions you only get half that consistently while the MP-3's produce 640 ppu's consistently with no problems and no headaches from the operators who use the gear in the field. In fact the MP-5's are kept on the lot and are rarely used, the maintenance guys will get pissed if you even ask for it.....and why would you when the MP-3's get the job done 99% of the time with no stress from power loss? -The moral of the story is power means nothing if its hard to harness or not OPTIMAL for its intended purpose.
They actually got rid the Hamster in the wheel and replace it with a smaller chip. So now the Ps3 is supposed to be easier to develop for? Not! what a great way to make existing owners look stupid.
I remember when the PS3 was originally announced, it did not have a GPU. The Cell was so powerfull, it would not need one. So if it's so powerfull with graphics, why at the last minute did Sony purchase a GPU released in 2005? Also proves the point that the Cell was designed for certain tasks. The SPU's also cannot handle code like the PPE can. Why do you think all developers are saying AI programming is easier and the results are better on the 360? All code must be crammed into 2 threads and sent to the PPE, then the PPE tasks out to the SPE's for number crunching. Bottleneck-HO! Said it before, great for TV's and cable boxes, bad for gaming. Cue DJ with the "fake facts" counter...
before that this was a step towards the PS3 in my last comment. I meant this was a next step towards the PS4.
CELL is put it on today's computers considering how slow they are and how sucky the can really be theMart doesn't know crap about Sony on what there doing with the new CELL knowing that he's a major boner in everyway why don't you lay off the retardation pills ok theMart and oh by the way that just proves my point even more cause of your retarded username.
"Why do you think all developers are saying AI programming is easier and the results are better on the 360?" Have any published articles or interviews deliberately stating this claim? You won't find it, I guarentee. "All code must be crammed into 2 threads and sent to the PPE, then the PPE tasks out to the SPE's for number crunching. Bottleneck-HO!" That's completely false because it's simply A method of programming the Cell, not THE method. This method is used because the PPE is the most familiar architecture in the Cell CPU, and thus is used as the glue to hold everything together. There are other programming models that are starting to appear where the SPEs and PPE are treated as independent processors (yeah, they're independent to begin with) and stream data to and from each other rather than trying to shove everything back into the PPE. It can even be programmed to run like the Xenon CPU, albeit with more on-chip memory and faster bandwidth. The ultimate programming method (which is the goal of Cell programming) is completely ignoring the PPE and just running everything on the SPEs. The reason they want to do this is because G5 core/PPE utilizes L2 Cache. But wait, the Xenon only has L2 cache, so what's wrong with cache? Cache is basically random memory. Cores that utlize L2 cache have no idea what's in that on-chip memory, and if they ask for something that's not in L2 cache it takes hundreds of clock cycles to search for it in RAM. Cell's SPE architecture completely eliminates that performance hinderence by having completely control over their Local Storage. It's just as fast as cache, but with more benefits. If there's a data miss in the SPEs, it only takes Cell around 8 clock cycles to find it vs. up to 200~300 clock cycles if it's missing from the G5 core. These data misses happen more often than you think, and add up to a lot of core downtime on traditional CPUs such as Xenon. Really, the big advantage of Cell is that it gets a lot of performance by eliminating big problems that programmers have had to suffer through for a long, long time.
wrong with the old Cell processor?
N4G is a community of gamers posting and discussing the latest game news. It’s part of NewsBoiler, a network of social news sites covering today’s pop culture.