Sony PS3: Doom & Gloom → Heh... cutbacks

Everyone whined about the shortage of 360's and said the PS3 will be better at launch... well it looks like its heading down the same road..

Time will tell.

Ryan

Technicaly it's better than the XBOX launch, but hardly. People who want a PS3 will either be camping out on launch night or waiting a while.
I think the important thing for Sony is that they have enough for holiday sales.

agreed.

i camped for my 360 and found it to do the 3 red lights of doom a week later..

and rumor has it the PS3 is being shipped where 1 or more cores on their 8 core CPU is dead. However the PS3 only requires that 7 cores be operational for it to boot.

Ryan

So, 1 or more dead, but requires 7 to boot, makes mathematical sense

Um ... where have you heard that? Link to anyone reporting it?

You know those silly rumors, just so much info on the net

»www.dailytech.com/articl ··· sid=3295

That's ABSURD, why a console would need 8 cores is beyond me to begin with, then 10-20 percent that actually work? Wow, that's just.....shakes head. It's a shocker the price for the PS3 isn't higher due to just THAT fact alone.

Edit due to spelling, in bold. Thanks for your cooperation in this critical matter.

That's not really rumor, it's pretty widely understood. The specification for a cell chip is to have 8 SPE's and 1 PPE. By default, one of the SPE's is going to come turned off. Developers have known all along that they are only going to have 7 SPE's and 1 PPE the whole time, so this really isn't an issue.

From the article:

-- With Reeves' statements, it appears as though Sony will be producing PlayStation 3 consoles with different Cell processors -- some with all eight cores operational and some with just seven. Reeves however does not believe that Sony will offer different pricing for the machines and only time will tell if there will be performance differences. Reeves says however that users will not see any differences in speed. "The PlayStation 3 only uses seven of [the eight cores]. You'd have a spare."--

I don't see a problem if IBM is already saying the PS3 will fully function with no degradation at 7 cells. The thing to look out for is the possibility of a second cell burning out and making your PS3 an expensive deadweight (after warranty). Doesn't seem like it would be likely with the testing IBM is doing, and even if there is a percentage out there Sony will likely do the same thing Microsoft did with the Xbox (DVD drive scratching disks to the point they are unreadable) and just keep repairing or replacing the faulty equipment regardless of warranty .

Yeah, it's pretty much a non-issue. While most of the processors made will not have all the cells working, the PS3 will be shipping with chips that have either 7 or 8 cells working. There is no reason (yet) to believe these cells have high burnout rates. If I were to buy a PS3 I would wait before buying to see what happens.
It's a further non-issue for me as I'm going for the cheap-o wii.

It's just a cpu, Cpu's don't generally go bad. Besides I seriously doubt that the chip will be able to function if a cell burns out whether it has 7 or 8 working to begin with. I imagine the chip goes thru a grading process and after that process something is done to lock in the number of cells the chip will use be able to use; similar to how many of the first PII chips were made into Celerons.

From what I've read previously about cell, if Sony is shipping any PS3's with 7 cell's developers will only develop titles that will use 7 cells. It doesn't seem that it's a matter of throwing code at the processor and it figures out how to distribute it amongst the cells; the compiler generates code that say's "I need six cells and here's what they'll do" thus they'll have to compile the code that they write for the lowest common denominator, that being seven cells. Furthermore most of the games at launch won't even use the cells for much relying on the PPC core to do the work. Again this is what I get from various things I've read over the past year or so, correct me if I'm wrong.

That DailyTech story seems fishy if you didn't notice.

iirc, even if a cell burns out, the rest would remain intact and function. they all connect to a central PPE module that dictates where specific instructions go to, which SPE (one of those "cores") will do the work, and which are free to assign other tasks. So if one is dead. it will just skip it and go to the next available core. It should anyway.

As far as the possibility of cores burning out. Its hard to say. this is a new technology, brand new. We dont know for sure if the architecture itself tends to become overheated, if the internal stores (think l1 cache for each of the cores) tend to overheat. Or if the whole process itself will tend to be bottlenecked by some performance problem in the PPE (which is the single most important thing needed to funtcion in a cell chip)

while im sure the developers CAN write code specifically to use a number of cells, im pretty sure they can also just wring code segments that will be queued and run in order of available spe.