Is the Cell a FrankenChip?

[standing ovation]

You see, but that's the point exactly: to insource what has traditionally been a very outsource heavy aspect of Sony's CE operations, thereby increasing profitability (assuming high utilization rates at the fabs).

You don't need optimal production -- or even an ideal product -- to be profitable.

Success begins with the consumer. And without pandemic success, the cost of homegrown silicon will far exceed its return.

(Q) So what would prompt cash-strapped Sony to jump into a cesspool teeming with wannabe Intels?

(A) That Toshiba and IBM also have multiple Cell factories should say something ...

As for x86 - yes it fixes problems - but primarily targeted at media and imaging it seems, and other FP-intensive ops. Never billed as a replacement for that architecture by any means.

Well, duh! Only a fool would broadcast their itinerary before engaging the enemy.

[size=-2]"Beating Intel at its own game isn't a trivial undertaking, as companies such as Motorola, Cyrix, and Transmeta have all learned." -- Electronics Design Chain Magazine[/size]

 

[Carl B]

Standing Ovation, I don't know - maybe I've been misreading the tenor of your posts up until this point because that last response threw me for a loop.

I still think a fab build-out, pandemic success aside, makes perfect sense as long as the present and suspected-future need is there. And the need presently *is* there, in the form of Cell, RSX, and other chips for various electronic goods (from the semiconductor slides, seemingly primarily for imaging devices).

I see now that you're actually a believer in the shoot-the-moon strategy of Cell rather than a doubter (unless I misread your tone again), but I myself stay middle of the road. It has a moderate list of things at which we can reasonably expect it to succeed at, and which will have made the project worthwhile, as well as another list of potentially paradigm-shifting features with a much lower chance of coming to fruition in a competetive market, but nonetheless some chance still. And of course also a small chance of being a 'failure' in the unlikely event that PS3 also 'fails.'

So a number of scenarios, but for my part I feel positive on Sony's build-out decisions to this point. On your end I find you slippery to pin down in terms of what you think is good and what is bad, but either way I think I'm understanding that you're expecting *BIG* things, whether they be positive or negative.

But I will say with regard to your angle taken on the costs of these goods and their profitability, if Sony can fully utilize fab capacity, in theory that should add to the profitability of their CE division by reducing dependence on outsourced chips. I think you place too much emphasis on consumer/market success - margins play an extremely large a role in determining end profitability. There's something to be said for outsourcing vs insourcing solutions in a number of situations, but Sony is frankly large enough that vertical integration - I feel - makes sense in areas they consider themselves expert in.

The slides explain their motives and thinking as well as anything I can post, and I think they bare out that Sony is interested in pursuing increased fab capacity and advanced processes for reasons stemming from - but going beyond - the Cell processor.

Remember that on the global fab scene Sony is gaining more and more muscle, to the extent that they are already starting to brush up against TSMC in terms of sales volume.

 

[standing ovation]

I still think a fab build-out, pandemic success aside, makes perfect sense as long as the present and suspected-future need is there.

No matter how rosy the prospect may seem, this is the reality: getting into a market that isn't very profitable is one of the better ways to fail.

It is like entering an Olympic event with hopes of taking gold but having never played the sport. So long as there are seasoned competitors out there (engaged in a similar activity), victory will be a long shot.

I see now that you're actually a believer in the shoot-the-moon strategy of Cell rather than a doubter (unless I misread your tone again), but I myself stay middle of the road. It has a moderate list of things at which we can reasonably expect it to succeed at ... And of course also a small chance of being a 'failure' in the unlikely event that PS3 also 'fails.'

Sony and Toshiba (and to a certain extent IBM) are preparing to storm the market because they have got a Renaissance up their sleeve. Not only is the design technologically superior to competitors' offerings, but so is its manufacture and operation. Cell plants are springing up like weeds because the trio has plans to peddle its wares -- the chips, its manufacturing process, and even their operation (by selling processing time).

In other words, STI will begin to propagate a new standard which they author and control. So Cell's utility outside of the entertainment sector should eclipse its usefulness within it.

 

[Carl B]

No matter how rosy the prospect may seem, this is the reality: getting into a market that isn't very profitable is one of the better ways to fail.

It is like entering an Olympic event with hopes of taking gold but having never played the sport. So long as there are seasoned competitors out there (engaged in a similar activity), victory will be a long shot.

We'll just have to agree to disagree here. I don't see Sony as 'getting into the market' so much as they are in the industry to fulfill an internal demand. Not like they're trying to compete with TSMC or Chartered here or anything. In that sense I think that the move towards these fab build-outs is a low-risk 'positive' for them regardless.

On the side I'm kind of interested in that Japanese chip consortium that's forming which Sony is remaining on the outside of; wonder if they feel their traditional Toshiba and IBM alliances are enough on that front?

Sony and Toshiba (and to a certain extent IBM) are preparing to storm the market because they have got a Renaissance up their sleeve. Not only is the design technologically superior to competitors' offerings, but so is its manufacture and operation. Cell plants are springing up like weeds because the trio has plans to peddle its wares -- the chips, its manufacturing process, and even their operation (by selling processing time).

In other words, STI will begin to propagate a new standard which they author and control. So Cell's utility outside of the entertainment sector should eclipse its usefulness within it.

Well, I think that's definitely a goal being pursued, and an ideal scenario for STI, but I just think it's too early in the game to really be addressing the 'Cell Everywhere' strategy. I mean hell Cell roll-out hasn't even begun yet and competetion is going to be fierce on a number of fronts, but I like your confidence!

 

[standing ovation]

We'll just have to agree to disagree here. I don't see Sony as 'getting into the market' so much as they are in the industry to fulfill an internal demand. Not like they're trying to compete with TSMC or Chartered here or anything.

This is exactly what they are doing!

Otherwise, it would make more cents to outsource their designs (like Apple, NVIDIA, Nintendo, ...) and let somebody else worry about that headache. Scribbling a bunch of gates on a piece of paper is relatively easy, but making it a reality -- in volume -- is where the difficulty increases exponentially.

Well, I think that's definitely a goal being pursued, and an ideal scenario for STI, but I just think it's too early in the game to really be addressing the 'Cell Everywhere' strategy.

Why not? IBM and Mercury Computer Systems seem to be doing that right now ...

 

[winterbird@nerdshack.com]

Otherwise, it would make more cents

sense....

 

[rounin]

It was a pun

 

[typoEDR]

I bet he registered just to post that.