Does Cell Have Any Other Advantages Over XCPU Other Than FLOPS?

[MrSingh]

Korean and Chinese companies will own the future market of CE devices (and every segment that Cell wants to be in), and Cell isn't going to change anything for Sony. (other than being a damn fine CPU for a game console)

 

[Vince]

and Cell isn't going to change anything for Sony.

How can you say this when Sony's Semiconductor Group has already shown massive change in what is just the prologue to what some of us propose. In 2005 Sony had the highest growth of any top 20 semiconductor manufacturer at neartly 20% growth. This growth was based entirely on usage within Sony Group, external transfers declined while internal Group transfers jumped by 50%.

You can't say it's not going to change anything when the change is already happening.

Korean and Chinese companies will own the future market of CE devices (and every segment that Cell wants to be in)

Untill the day that the "Korean and Chinese companies" (taken as the commodity manufacturers) have R&D budgets on the scale of a TI or Sony, they will never own the market. Take HDTV; Sony alone owns 50% of the microdisplay market and 20% of the LCD market.

This was by way of SXRD and 3LCD sales, both proprietary. I don't see the 3rd world commodity manufacturers competing in the moderate to high-end TV marketplace which requires either massive upfront research/development expenses (eg. SXRD) or going 3rd party with what is the equivalent of putting Cell in a TV, buying DLP from TI.

 

[expletive]

A term: "Product differentiation" ... lost on Aaron. In a time of increasing commoditization, one needs to find a way to differentiate himself; technology is ultimately the way it's going to be done. You just helped make the case for my position, thank you.

Using a Cell processor, regardless of the cost does not guarentee product differentiation or even 'best of breed' wahtsoever. Its very likely that the best mass-market HD deinterlacer chip will come from someone other than Sony (think DCDi technology, which imo is still the best mass market SD video-source deinterlacer available). Then what? Sony is joined at the hip with Cell and whatever software theyve written for it. Now Panasonic and Samsung are off buying these chips for a pittance and getting a competitive advantage as well.

I think this is overlooked somewhat in that if the STI cell-based solution isnt the best one available theyre stuck. They can no longer choose best-of-breed, they have to rely on making it for everything they plan on using cell in.

 

[Vince]

Using a Cell processor, regardless of the cost does not guarentee product differentiation or even 'best of breed' wahtsoever.

Again, will people stop taking what's stated and mentally inflating it into this all-or-nothing, best/mythical/legendary/only mentality?

I never said that; why people post such things is beyond me. Is it like you feel compelled to make sure every last soul out there knows that Cell isn't the only solution? Which nobody stated to begin with... I know this is how Aaron thinks and it befuddles me, but why? Good things can be posted about Cell without a needed counter-argument downplaing it... really.

You guys remind me of immature teenage boys talking to a girl; for every compliment there is an immediate and automatic overt compliment back; reciprocation where none is necessary. 'You look nice... Ohh, you look nice too' -- Seriously, grow up.

EDIT: Actually, it's like two teenage boys talking about their birthday gifts; ever escalating in intensity and importance: Yeah, but mine has this and that... yours can't do this. Uhh, great. That doesn't change what I have, I like it, it works well, so shut the hell up already.


I was merely talking in terms of enabling such differentiation; Cell is an enabler, ASICs and DSPs are not. It's just that simple. Nobody stated that Cell was going to replace x86, Sony was going to amazingly surpass Intel, that Cell would be the last IC ever designed or used, or even kill Alpha... again.

 

[MrSingh]

classic. I was wondering how long before he'd start throwing insults.

don't ever change, we love you. （＾3＾）－☆ちゅ！！

 

[flf]

I was merely talking in terms of enabling such differentiation; Cell is an enabler, ASICs and DSPs are not. It's just that simple.

Can you give some hypothetical examples where CELL is such an enabler, and a custom ASIC or DSP somehow can not fill the same role?

 

[patsu]

So to break it down there is no lack of processing power in CE devices right now, just lack of good ideas?

That's basically what you're saying right?

Pretty much. All the enhancements people have come up with fit comfortably within the existing processing envelope. The primary driver in the CE space always has been cost, with the secondary driver being cost, with the tertiary driver being cost, eventually features come into it, but features are at best a temporary advantage often lasting less than 6 months.

Remember that we are talking about devices that in mass production have a BOM of under $50.

Aaron, I actually believe in the above statements. But I take issue with the way it's put across. I see them as obstacles, not laws (as in laws of physics).

Is it possible that with the right ideas, the cost factor can be mitigated ? Think Apple.
Also our view of the CE space may change in the future as brand name device makers start to build hooks into their boxes, like Xbox Live and iTunes.

Secondly, and this is a "thinking out loud" question: As digital content catches on, would the content providers including Sony be interested to invest, control and deploy a pervasive, strong security CE platform ? Something locked deep inside a processor, that can be upgraded safely and on-demand (cannot rely on hardwired logic), plus uniform so that it can be dropped onto all content devices (reduce weakest links in security)... various incarnation of cells perhaps ? Is this possible ?

Finally in a cheating sorta ways, is it possible for another device to make use of the Cell inside PS3. So one does not need to actually deploy it. Then Sony charges 'em the Cell tax like Apple charges for iPod tax. One of those "what hair-do fits me best" thing with EyeToy and Estee Lauder (or whatever). I always want to know how short I should cut my hair (once with disaestrous result, so I don't try anymore 8^D ).

It's common in the PC world (where technically digital cameras rely on PCs to process and store photos). It won't replace PCs, but just like everything else... some prefer Mac, some Windows, Linux and others Amiga. Plus sometimes the PC is just too cranky/geeky for important things like hairdo 8^)

 

[avaya]

Can you give some hypothetical examples where CELL is such an enabler, and a custom ASIC or DSP somehow can not fill the same role?

Marketing.

"This Sony/Toshiba HDTV uses Cell technology, same technology that saves lives, guides missiles...." etc.

It's all about the marketing. This is the point missed by so many.

Sony’s whole strategy now is to fight commoditisation. A custom ASIC may do the same things but is it going to carry the same marketing weight? That's the whole point.

 

[Vince]

Can you give some hypothetical examples where CELL is such an enabler, and a custom ASIC or DSP somehow can not fill the same role?

And on your comments of "dedicated chips," it's related to my comment to Aaron on implications of NFLT in computing; DSPs are only useful for their specific and static program landscape; they are useless outside of it. NFLT dictates that you look to the niche and back-propogate for the strategy which is optimal if you want to be sucessful. Cell is dynamic and yeilds a balance between the preformance of a DSP and the general computation of a CPU, targeting the digital media computational landscape. They looked at the near future needs and designed an architecture around it: HD stream manipulation; multiple HD stream decoding; HD recording, compression, archiving, searching; DVD archiving and HD upconversion; Video-over-IP to client devices; augumented reality, visual-based GUIs; Multiple Tasking, IM, SMS, video email; Seemless device connectivity; etc.

You will not find another system, especially a static DSP based on which can come close to this type of functionality, across such a broad spectrum, so effortlessly.

I haven't seen a consumer DSP do the following. More importantly, I don't know of an ASIC which can do all of the above mentioned features:

Toshiba brought 2 demos for this Cell devkit, well-known 48 MPEG2 streams playback and the new "Digital Kagami Type F" (digital mirror type F) demo. This new image-based rendering demo converts user's face into polygons and applies makeup to it by texture-mapping, template-matching and motion prediction by matrix calculation, and makes it into 720x480@30fps in realtime, on a half-mirror.

Trying on different makeup styles using the "3D makeup simulation"​

 

[mckmas8808]

. Nobody stated that Cell was going to replace x86, Sony was going to amazingly surpass Intel, that Cell would be the last IC ever designed or used, or even kill Alpha... again.

OMG! Somebody finally said it. Some people feel like they have to fight to their death to prove that CELL is only good for the PS3 and for some reason Mercury Systems. Why is it that Mercury Systems is the only company that would want the CELL processor? I'm thinking that it might be possible for the CELL processor to make CE products better. For people to put that thought down and say it never going happen is living in the land of OZ.

 

[expletive]

I haven't seen a consumer DSP do the following. More importantly, I don't know of an ASIC which can do all of the above mentioned features:

Trying on different makeup styles using the "3D makeup simulation"​

Cool, as soon as the mass market embraces 3d makeup simulators, Sony is all set.

 

[mckmas8808]

Cool, as soon as the mass market embraces 3d makeup simulators, Sony is all set.

The point is it can do things that others can't.

 

[Vince]

Cool, as soon as the mass market embraces 3d makeup simulators, Sony is all set.

As I'm sure we both know, the underlying technology is not limited to just implimentation in 3D make-up simulation... Dr. Richard Marks has shown Cell powered 3D GUIs and Augmented Reality type uses with the wireless HD IP camera.

I can't imagine its that far fetched to have such a system which recognizes your facial features (as the Toshiba video did 500 points) and then loads personal things: Say your saved DVR list, TV channels arranged by your viewing habits or shows/movies which would likely interest you, perhaps IM, Email, SMS. Hell, they could pull a Google and sell ad space tailored to you, ala. Minority Report.

 

[expletive]

OMG! Somebody finally said it. Some people feel like they have to fight to their death to prove that CELL is only good for the PS3 and for some reason Mercury Systems. Why is it that Mercury Systems is the only company that would want the CELL processor? I'm thinking that it might be possible for the CELL processor to make CE products better. For people to put that thought down and say it never going happen is living in the land of OZ.

No, i think what the debate is surrounding is the propagation of hype surrounding what the processor will mean to CE (and civilization as we know it ). Vince, and now you, want to turn this around so the arguments that dont see the cell as some sort of CE epiphany as 'bashers'. Theres just no compelling reasons right now, or at least none that have been presented, that convince me that CEs are now on a new, better, course since the use of Cell is imminent. Thats really all ive ever debated, with the counterpoint being something along the lines of "But now your TV can be upgraded from 12 HD streams to 24 HD streams decoded simultaneously". Also, theres usually not much discussion on costs and really what the people will want or buy. I think we tend to imagine a world where the best technology wins and thats not always the case. So ive tried to bring that perspective to the table as well.

I think the best argument is honestly the marketing one by avaya. I'm not convinced its all AS big a deal as he feels but theres certainly a chance. Trinitron was a household name and carried some wieght for awhile. I think trinitrons were visibly superior for awhile too so Cell will have to provide SOME tangible benefit along with the marketing imo.

 

[expletive]

I can't imagine its that far fetched to have such a system which recognizes your facial features (as the Toshiba video did 500 points) and then loads personal things: Say your saved DVR list, TV channels arranged by your viewing habits or shows/movies which would likely interest you, perhaps IM, Email, SMS. Hell, they could pull a Google and sell ad space tailored to you, ala. Minority Report.

Heck if we get to the point where we the consumers and the associated technology is ready for that, and the cell is a uniquely qualified component in providing it, i'm sold.

 

[mckmas8808]

Heck if we get to the point where we the consumers and the associated technology is ready for that, and the cell is a uniquely qualified component in providing it, i'm sold.

And THAT expletive is my and Vince's point the whole time. I and Vince have never made the CELL look like the holy grail of processors. Just that it might or should be able to give the consumer something that is not available to them now.

And that example of Vince's that even you say you're sold on is the point.

 

[Vince]

No, i think what the debate is surrounding is the propagation of hype surrounding what the processor will mean to CE (and civilization as we know it ). Vince, and now you, want to turn this around so the arguments that dont see the cell as some sort of CE epiphany as 'bashers'. Theres just no compelling reasons right now, or at least none that have been presented, that convince me that CEs are now on a new, better, course since the use of Cell is imminent. Thats really all ive ever debated

Great but, atleast I, really don't care; I'm sorry. If you're so intent on discussing it then I suggest you start another thread in the Hardware Forum and discuss DSPs in CE applications there; but in this thread we were discussing possible uses (eg. brainstorming) of Cell in CE applications.

This is a technology forum, we discuss technology; not marketplace economics. If we want to dream up potential uses then that's allowable concidering the thread we're in; just like it's fine to talk of how ATIs GPUs kickass in GPGPU work without having somone come in and talk about how nVidia's are better priced.

As I stated before, there is no need for your fallicious goal of balancing supposed hype: reciprocation where none is necessary. Several posters truely act like teenage boys talking about their birthday gifts; ever escalating in intensity and importance: Yeah, but mine has this and that... yours can't do this. Uhh, great. That doesn't change what I have, I like it, it works well, so stop downplaying it already...

 

[avaya]

I think the best argument is honestly the marketing one by avaya. I'm not convinced its all AS big a deal as he feels but theres certainly a chance. Trinitron was a household name and carried some wieght for awhile. I think trinitrons were visibly superior for awhile too so Cell will have to provide SOME tangible benefit along with the marketing imo.

Just to clarify I don't think it's a cert that it will provide the brand power necessary to achieve "like.no.other" status. Like you said it will have to provide some sort of real benefit, even if for a limited timeframe, for them to pull it off on the marketing side.

I'm not going to make predictions. Both sides on this "Cell in CE" debate have compelling arguments.

 

[Dave Baumann]

Again, will people stop taking what's stated and mentally inflating it into this all-or-nothing, best/mythical/legendary/only mentality?

Actually Vince, this is all because of the way that you convey things.

 

[Vince]

Actually Vince, this is all because of the way that you convey things.

Oh really, Dave? Can you quote where I stated such black and white statements concerning Cell. If I did, I'll be happy to edit it, so I'll be eagerly awaiting your reply.