Well, withought a doubt we will see delays, but in the end Sony will pull it off quite well Im sure of that, with titles like MGS4 and RE5 for it, not to mention numerous others. But as ProgrammingAce has pointed out, there is no way it will be finished in time for its release date, unless Sony wants to release the console reginally, as in Japan or U.S. first then Europe, that will be a mojor thorn in some peoples asses but it will allow Sony to make a good console.
Sony has already stated that all games should use the processors in a supervisor/worker relationship. Any other method would be folly. If you were to try to control 7 processes at the same time, half of the processors would be sitting idle waiting for the graphics to finish. According to sony, this is how it should work. The main core is the only one that can access memory directly (odd, but true). That core acts as a supervisor, splitting the workload into smaller chunks of roughly equal timeframes. So instead of splitting the workload by size, it gets split by time. That core then sends a pointer to the worker processors to calculate the requests and send it back to the main core where the workloads are reassembled and the frame is rendered (frame in this case is an instance, not a graphical frame. A frame would include the graphics, physics, movement, AI, etc, etc, etc) Overall, it's a hell of a system to keep straight. Sony basicly demands that all games be programmed this way because the weird memory access routenes aren't availible to the worker processors. I'd rather program assembly code on saturn then C++ on the PS3. EDIT: Damn... until just now i had always assumed that i'd be picking up a PS3 at launch. Now the more i think about it, i think i'll just wait...
Maybe they ll release a special version of C++ taking into consideration the memory bottleneck you just mentioned. On a side note, I does make sense to allow only the master processor to access the memory, since this would warrant a supervisor/worker relationship, act as a distribution point (tighter control) and avoid what they did on the saturn where the Sh2s had to interchange their access to the memory. Imagine what would happen if you d try to access memory from more than one core at the same time, the system would drag or mess up realistically. You should be afraid of coding assembly on the Ps3 more than C++ on PS3 actually
This reminds me of the debates on processor speed that were going on in the 16-bit days. Personally, I don't know why they stick with C++ when they could just make up their own language. Not only would companies be forced to pick an exclusive format for their games (I am so sick of seeing triple releases, what's the point of bothering with hardware?) but they might just be able to shed some lines of code and build tailor made engines that will utilize the hardware's strengths. Then again, what do I know?
Odd and a seemingly VERY poor architectural choice. The common data bus (or whatever Sony wants to call their version of it) is going to be a huge bottleneck with all of the workers competing for indirect memory access.
If its time-slot based there's no competition. Each unit has an allocated time-window to do or to not do.
a rare look into the PS3 processor The first core of the PS3 processor! http://www.edbott.com/weblog/images/hamster_wheel.jpg
Technically, they would still be competing as only one worker can use the memory access at the same time. Using separate cores with separate caches and memory ports would be a seemingly better choice. Of course, I have not seen the actual architectural spec for the PS3, so there may be a legitimate reason for the master-slave core memory relationship.
I imagine that the master/slave concept pimarily derives from the fact that games are traditionaly a master task that has sub-applications running, unlike computers that are aimed at equality of power accross all of the applications , since they can't truly decide the priority of one over the other. The ps3 just organizes the sub-applications of the main game code in a more efficient way, but this may very well limit and slow down things if not handled with caution at every step of the way
You're friends suck/have bad luck/need to learn how to take care of their systems. Mine, my coworkers, the unit at work & my friends units all work just fine. (all but the work unit are launch units too, mine went through dhl shipping to go from japan to me). :noooo:
am I the only one in the world who would love to code in this machine just b\c of cell processor? I think it would be a very intresting experience to code for such a machine. ^^
Don't get me wrong, I m as interested in coding this, as I was for the Emotion Engine and the Vu1 Vu0 setup. But a hobby-challenge is one thing, and making a functional product by a deadline is another.
I don't really think it's the PS2 at fault, but more a case of Sony having less tight quality control than Nintendo. Based on a recent and quite comprehensive showcase looking at the production of "Shadow" (which I forgotten the location of - sorry!), it seems as if the developer was pushing Sony's platform quite hard. Ico had its own problems, though I feel those were more to do with the fact it was a PS1 game carried over and completed in as little time as possible - much like The Getaway, for example - and perhaps the usual troubles associated with first generation releases as opposed to the actual system's performance. Of course, I could be just as wrong as the next person...
No, I think that it is perfect for offloading physics, AI, game state and other aspects of a game to individual dedicated processors. I just have a problem with how the memory is supposedly handled.
all this discussion about the Cell however.. almost makes us forget the rest of the system. Any news/info on the Audio and Graphics system?
I was kind of wondering that myself. I'm sure that there is some kind of separate sound/graphics processor, but from everything I've read, the Cell is the end-all be-all for this system. Is everything getting pushed through the master/slave or is there some offloading to an actual sound or graphics chip? And thats not even asking about memory allocations for them if there are....
Same here. Just wrapped up 60 hours of Oblivion progress too. Not a single hiccup or freeze to speak of. Launch day unit. Failure rate of the 360 is highly exaggerated.
My 360 is still up and running no over heats or anything it froze on me once when i dropped from the top of my tv but it started right back up ones i cleaned the disk. the one thing that could kill this forum is a fan boy for any side. I know there are alot of people that like ps3, 360 or Rev. But one thing we all have in common are games and that is the main purpose of the forum. It just pains me to hear people someone say that 360s are failling all over the place and killed my baby chickens (i knowassembler had problems) but 10 people in my company have them and no body has problems other than games are to fucking expensive. So less fanboy and more hard facts is all i want. I never liked sony/ or hardware but i won't buy first gen of any of their products unit 3rd rev. I love my slim line pstwo.