Cell 65nm production started

I never really have seen an explanation I can understand regarding why GPU's are clocked so much slower than CPU's. AMD/ATI merger going to change this or..........??
 
I never really have seen an explanation I can understand regarding why GPU's are clocked so much slower than CPU's. AMD/ATI merger going to change this or..........??

Think of it this way: GPUs are made from libraries of higher-level building blocks, while CPUs are assembled at a lower level, like programming in a high-level language vs. programming in assembler.
 
Anyone know how much is the excepted to save them? Either direct cost? Estimated cost? Yeild increase? I know it will help and I have heard others say it will help ALOT. Just wondering...
 
Anyone know how much is the excepted to save them? Either direct cost? Estimated cost? Yeild increase? I know it will help and I have heard others say it will help ALOT. Just wondering...

In theory, you could consider a node move from 90nm to 65nm to ultimately reduce the cost of a chip by 50%. But in truth it's more complicated than that. The potential savings though are obviously significant however; that's what needs to come through here.
 
My point is, Sony isn't going to roll out a new console revision every time a component is cost reduced.

They're not going to have PS3 Mk II with 65 nm Cell and PS3 Mk III with 65 nm RSX.

They are more likely to wait until both are shrunk before rolling out a Mk II, at least based on history.

But who knows, maybe they're getting so killed on costs and the need to push prices down that they will dribble out revisions incrementally.
 
My point is, Sony isn't going to roll out a new console revision every time a component is cost reduced.

They're not going to have PS3 Mk II with 65 nm Cell and PS3 Mk III with 65 nm RSX.

They are more likely to wait until both are shrunk before rolling out a Mk II, at least based on history.

But who knows, maybe they're getting so killed on costs and the need to push prices down that they will dribble out revisions incrementally.

Go to wikipedia (or a better source if you know of one) and look up the number of hardware revisions that took place on the PS2 and maybe you'll change your opinion. The PS2 has had around 14 revisions that averages out to about two a year.
 
is that why model number for ps2 had range from 10000 to 70000 since its launch???

Yup. One of the things people seem to be overlooking with the PS3 launch vs. the PS2 launch is that Japan saw two versions of the PS2 hardware before the PS2 was ever launched in the United States. Sony has had a rough start with PS3, but they didn't do any better with PS2, particularly.. it's just we in the United States were spared the true launch in Japan, and got a refined model awhile later.
 
Yup. One of the things people seem to be overlooking with the PS3 launch vs. the PS2 launch is that Japan saw two versions of the PS2 hardware before the PS2 was ever launched in the United States. Sony has had a rough start with PS3, but they didn't do any better with PS2, particularly.. it's just we in the United States were spared the true launch in Japan, and got a refined model awhile later.

it doesn't seem the hardware is having major issues like XB360. which is nice.
 
What major issues? It's not like there are 10 million PS3 sold anyway to even have problems crop up. ;)
usually its the launch units that are the most troublesome, and so far, there have been very few reports of ps3's failing for whatever reason.
 
Is it safe to say that RSX may be going into 65nm production soon as well?

Also, does anyone know the status of Sony's 45nm fab facility? How soon could we see Cell/RSX go into 45nm production?

(I know it's too early to think about that, but just for fun.)
 
Also, does anyone know the status of Sony's 45nm fab facility? How soon could we see Cell/RSX go into 45nm production?

(I know it's too early to think about that, but just for fun.)

45nm is expected to be available in 2009 according to Goto.
 
I totally expect them to start with such chips, which are relatively easy and would be tiny at that process, before venturing into much bigger and more complicated architectures like Cell BE.

Roughly if they have a scaling about 60% in each node it would be very tiny. I dont care to estimate a possible diesize (lazy :)) but someone else could do that hehe. ;)
 
Roughly if they have a scaling about 60% in each node it would be very tiny. I dont care to estimate a possible diesize (lazy :)) but someone else could do that hehe. ;)

It's quite amazing to see the size of the first EE and GS when they were at .23 (i think that's what they started as? or .25?) and then see how big they are now...
 
http://www.bloomberg.com/apps/news?pid=20601101&sid=aXr.PBbgNC6Y&refer=japan

Bloomberg said:
Any price cut won't affect Tokyo-based Sony's target to break even in the games division next fiscal year, Takao Yuhara, head of investor relations, said yesterday. He said the company is also looking to boost output and cut production costs to erase losses in the year starting April 1.

Is this a hint? Is April 1st the date they intend to start shipping units with 65nm parts? It seems to tie in nicely with the fact that 65nm Cell production is already underway.
 
http://www.bloomberg.com/apps/news?pid=20601101&sid=aXr.PBbgNC6Y&refer=japan

Is this a hint? Is April 1st the date they intend to start shipping units with 65nm parts? It seems to tie in nicely with the fact that 65nm Cell production is already underway.
Actually I think it ties in with the 6 million PS3s that they are supposed to have manufactured by the end of March.

Ken Kutaragi claimed last summer that they had secured component supply for the 6 million units they promised to deliver within this financial year.

That could mean they had secured a delivery of 6 million Cells at 90 nm and expected the 65 nm production line to be ready to supply Cells at a sufficient rate by March.
 
Back
Top