Cell and RSX : their relationship.
- Thread starter !eVo!-X Ant UK
- Start date
- Status
MechanizedDeath
Regular
When you have no proof, lie/stall/deflect. I see we have some masters of this dark art. 
With lead times, you can't base your design on the competition. Not without delaying, and it would be more than 6 months. Sorry, that's a pretty lame excuse there.
And I like KK. He dreams big, which is what you need when setting goals. If you miss the mark, that's fine, as long as you don't set targets that are so far away you lose sight of them. If you set easy targets you will always underperform. The project team I worked for at IBM had a great vision and was trying to beat Intel to the 1GHz target by designing a system from the outside in. The joke was that everyone was probably gonna be working back with S/390 in a few years...many internal projects get scrapped like that. But you have to set a high watermark, or what's the point in even trying? PEACE.
With lead times, you can't base your design on the competition. Not without delaying, and it would be more than 6 months. Sorry, that's a pretty lame excuse there.
And I like KK. He dreams big, which is what you need when setting goals. If you miss the mark, that's fine, as long as you don't set targets that are so far away you lose sight of them. If you set easy targets you will always underperform. The project team I worked for at IBM had a great vision and was trying to beat Intel to the 1GHz target by designing a system from the outside in. The joke was that everyone was probably gonna be working back with S/390 in a few years...many internal projects get scrapped like that. But you have to set a high watermark, or what's the point in even trying? PEACE.
well sony would have had a decent idea of what ms would get from ati for awhile now . Its not hard to see how the pc space has been evolving since the dx 8 era . So itcould be that knowing the power of a part nvidia and ati could provide sony knew from 2 years ago that the toshiba part would't be powerfull enough and then proposed changes didn't sit will with them .
I duno to much we don't know to base anything on . However I'm sure sony knew weather or not the toshiba gpu would be powerfull enough to compete against the pc parts for awhile now
I duno to much we don't know to base anything on . However I'm sure sony knew weather or not the toshiba gpu would be powerfull enough to compete against the pc parts for awhile now
I'm sure we'd all LOVE to see that Toshiba part. I'm guessing it might be that part referenced in the Cell (Superchips?) conference showing.
Not to my knowledge it wasn't.PCEngine said:
Yes because everything in this world is result of a Microsoft action. It's entirely impossible there could be any other reason for Toshiba GPU to be useless (say, maybe not being economical to manufacture), it's all because of "power", because after all that's the only variable in chip design right?
Yes and that embodiment required a MCM 1TFLOPS CELL. Go read one's posts on the history of PS3 where this is mentioned. Their goal was 1TFLOPS CELL.
Why do you think SONY rushed to get PS3 out? If MS hadn't decided to release Xbox 360 later this year, you wouldn't see PS3 till maybe 2007 where they can put a 1TFLOPS CELL 65nm chip into PS3. Some people are in serious denial.
PC-Engine said:
Thanks for the info. We shall see how close Sony and their partners get to their vision this time.
Why do you think SONY rushed to get PS3 out? If MS hadn't decided to release Xbox 360 later this year, you wouldn't see PS3 till maybe 2007 where they can put a 1TFLOPS CELL 65nm chip into PS3. Some people are in serious denial.
MS has to rush Xbox 360 out partly because their contract with nVidia will expire this year. PS 1 pre-dates PS 2 by 5+ years (counting Japan launches). Since PS 2 was launched in early 2000 in Japan, deliverying PS 3 in 2006 would be "normal" although PS Two is likely to continue selling at a reduced price (just like PS One).
I'm not familiar with Sony's plan and so I can't comment on whether 2007/2008 is the original PS 3 roll out date or not. I remember both Sony and Nintendo objected to launching nextgen too early in 2003/2004 when nextgen talk surfaced. In general, you want to prevent the Osborne Effect.
PC-Engine. I'd be bored to death if I drag this discussion further than necessary. So all the best to your future posts.
PC-Engine said:
If you want to get all technical about it. The Cell at 3.2ghz in PS3 is 216gflops. Yet that isn't the only form of the Cell. They could pull off a 4.6ghz Cell with all SPEs active and hit around ~350gflops. I'd say they did pretty damn good. You can make fun of them for not reaching their rather impossible goal of 1tflop (when there isn't a CPU around that gets even close), but it only makes you look rather silly. I'm still not sure if you are trying to imply that we should somehow feel slighted because the Cell is only a ~1/5th of their original goal in Gflops (PS3 only of course, other Cells could get up to ~350 gflops)? only 216gflops -- damn you sony/ibm/toshiba!
In addition -- since when is PS3 considered rushed? this is around the time that the new consoles should be coming out. I don't see how following the pattern of the past 20 (10 years for Sony now) years somehow means they are rushing it. I'd love to hear where you picked up that little morsel.
Haven't we been through this 100 times before?
The 1Tflop 4-core Cell chip was never meant to go on PS3. Unless PS3 were to come out in 2008. End of story.
The 1Tflop 4-core Cell chip was never meant to go on PS3. Unless PS3 were to come out in 2008. End of story.
Yeah and the yields on those chips are probably around 30%.
Regardless you're missing the whole point. SONY wanted to use a CELL based GPU called the Visualizer, but ended up using a PC GPU. SONY wanted a 1TFLOPS CELL Broadband Engine CPU in PS3, but had to settle for a 215GFLOPS CPU instead. They missed their goals by a long shot on both fronts.
- Status