Sony's Cell against Intel's Conroe

Once the Unix kernal has been ported for the cell will we then see a contender to the Conroe. Its probably gonna be a while until AMD gets it ducks in a row and I see this as a prime opportunity for the Cell to get its' foot in the door, so to speak.

The really great thing is Sony already has an established pc and notebook market with the VAIOs. Sony already writes its own software and there will certainly be software architects porting their own programs over to the cell.

Finally we should see some real magic in the marketplace. Unix and Linux will probably gain share again Windows, and AMD will have a friend (although Sony will still be a competitor) against Intel. But I believe the irony of all this will be Apple sneaking in the backdoor with a killer Mac machine running the Cell at its heart, put AMD out of business and cause Sony to kill the VAIO line of computers.

----------------------------------------------------------------------------
Frankly my dear, I don't give a damn! - Humphrey Bogart

All architechture arguments aside......

porting UNIX to cell will have no impact on Conroe which lives in the relm of Windows.

it would be competing in environments that use systems that hold many (>8) Power chips, Xeons and Opterons, Sparc. and it also would competing with only the segment of that market that uses UNIX.

Porting Windows to cell isnt going to happen anytime soon, MS wont waste money on that unless it translated to increased revenue somehow, and intel and amd are already serving the market quite well with proven platforms supporting 1 instruction set - x86.

Sony is not going to put Cell in their laptops and desktops running UNIX, noone is going to buy enough that to make it financially viable. Consumers will ignore it completely, look how many ppl havent switched to OSX and thats alot easier to switch to. Sony is also not going to sink money into building thier own OS to run on such a system, look at Linux, thats already maturing, and still its ignored in the mainstream.

I'm sorry i just cant continue, u have to be joking, please tell me u're joking.

Do u ever wonder if there is some sort of ""insane, spur of the moment, but incredibly wrong"" thought filter in our brains but that somehow it gets bypassed for some ppl??

I'm starting to think so. Shame we can do firmware updates or anything like that.

I think he probably thought it was briliant, probably still does.........

but he has to agree he pulled that completely out of his ass

Seems like he posted a random thought rather than a request for discussion or information.

The biggest non-PS3 customer are medical hardware companies so far. And that's not a very large market.

You're probably right; I forgot to mention that Cell was probably targeted for a niche market, yet in the article, they said it had other uses:

Why you would ever need to process that many different video streams on one TV set at the same time is beyond me.

Probably other *NIX OS's they would be the easiest to port it they have linux for it already. Also i am sure they would have a proprietry OS developed for it also, for specialized application.

If you look at it as the heart of a super home entertainment system (not unlike XBOX 360 ) it makes more sense to have a proprietry OS that wont be tampered with and just needs to do one set a functions till it stops working.

Frankly my dear Scarlet, I don't give a damn, either.
Clark Gable


Just kidding. I was wondering of the Cell would get anywhere as a PC architecture. It's actually the only way to compete with Windows. A Cell based machine wouldn't be "IBM-PC compatible" so there would be no "Windows Tax."


I wouldn't count on it happening soon but.....

Pffft who needs hard-drives when you've got cell!

Wait, you mean the registers and cache don't hold all the data cell needs?? Dang.

Isn't it the yellow thing?

Of course it can, its the cell! It can do anything!

That's pretty far for it to travel to get out to memory. Anyways the pic is heaps old.

Definitely, that board was an early reference at best, and I douibt ever saw/will see the light of day.

unless Cell can run Windows, its destined to be a high-load, calculating machine or just a really task oriented server. Be nifty to see a 8-way K8L Quad-Core Opteron running and controlling a farm of Cells designated for media, file serving, whatever your lil PC hearts desired, imagine the throughput

Where are you? are you going to reply to the posts here or did you just drop a random thought and never looked back?

Where are you? are you going to reply to the posts here or did you just drop a random thought and never looked back?
You know, a lot of people have been doing that lately - there like hit & runs, purpitrated by clowns...

I think I need to join you.

Thanks for the laugh. The Cell processor could not come close to competing with a K8 based Athlon, let alone a Conroe.

Yeah but they'll have general purpose cores that don't suck, better SPE equivilants and probably more specialised parts in 2010-2015.

In SP FP, yeah. Everything else, well I doubt it.

According to this chart... Cell does mpeg encoding 12x faster than a dual core G5.
(the site is french.. but you can still read the picture)

http://www.pcinpact.com/actu/news/ [...] ur-la-.htm


It all might be theoretical i donno. I can't find any 3rd party benchmarks.

last time i checked cell was for grapical power..not mathimatical...its math calculations (ie physics and what not) are up to par with a Pentium 3

it was an article in toms but it was mabey 1-2 years old if some 1 wants to dig it up.

It is all theoretical, and don't expect to see it available for 3rd party benchmarks, ever; the cell will stay a PS3 exclusive.

Either way, if that performance is even half as good as they say
i would expect this processor to find its way into the next G6 apple.

In terms of regular pc tasks however it won't be anything spectacular.

The cell may have an upper hand compared to the x86 chips, but seems like a poor replacment for using a dedicated graphics card. For example, you can do 3D FFT's directly on modern graphics card for a significant speed increase. I'd rather have something like a Torrenza platform with a high performance graphically intensive chip.

Odd that they mention 4D visualization in the article since thats more pretty images than anything diagnostic. I think memory problems are a larger issue since a 4D data set could be 1024x1024x1024 with 60 frames of 12bit data that has to be processed into 2D for visualization.