Does the Cell processor have a chance?

seebs

Newcomer
It might be. One of my coworkers dealt with them in another capacity once, and apparently they tend to blow off anyone who isn't likely to directly buy a LOT of hardware. I figure there's no reason for them to check that, out of a hundred people who said "I want to write about this", one particular guy might be a moderately successful writer whose articles might get read, when most of them are just dead blogs. :)

Still, it's sort of a shame. I really want one of those to mess around with. What Cell programming I've done has been neat, but I'd rather have a blade with real memory than a PS3 with 6 available SPEs and barely over 200MB to play with.
 

Nite_Hawk

Veteran
It might be. One of my coworkers dealt with them in another capacity once, and apparently they tend to blow off anyone who isn't likely to directly buy a LOT of hardware. I figure there's no reason for them to check that, out of a hundred people who said "I want to write about this", one particular guy might be a moderately successful writer whose articles might get read, when most of them are just dead blogs. :)

Still, it's sort of a shame. I really want one of those to mess around with. What Cell programming I've done has been neat, but I'd rather have a blade with real memory than a PS3 with 6 available SPEs and barely over 200MB to play with.

That's pretty much our problem too. We do have people doing development on PS3s, but it's even more of a niche than cell development in general. At least with a cell blade we'd have a small chance of getting it in our data center and making it a general resource for MSI users. There's no chance of that with PS3s.

Nite_Hawk
 

patsu

Legend
Where are you guys located ? I know of institutions with donated Cell Blades to encourage R&D activities.

EDIT: Oh... in Minnesota. Have you approach the schools for some value exchange (write about their programs in exchange for use of Cell and whoever are working on the Cell) ? I also know of an oversea location that allow companies to use their grid network and Cell blades for free (Some strings attached).
 

seebs

Newcomer
I'm in Minnesota, just ilke it says in the post. :)

The thing is, I'm not an "institution". I'm some guy. If I got a Cell system, it'd probably be in the basement about five or ten feet from the dryer. This is not an environment conducive to sales people drooling over the future sales prospects. :)
 

ADEX

Newcomer
Weird, I posted in this thread the other day but my post appears not to have made it...


Anyway, I think you've hit the nail on the head. IBM Mercury don't sell stuff to end users, they only sell to other big corps. This looks bad for end users but if they're selling Cell it is a good thing as it gives them a base to work from.

The first Cell was only really for the PS3 so it's not surprising they're not pushing it elsewhere much (other than HPC where it fits nicely).

The second gen should change things, the real HPC chip will come along and will appear in blades and possibly even workstations, but this is IBM so they wont be cheap...

Toshiba's Spurs chip is probably more interesting for end users as they appear to want to put them in laptops. If they sell then maybe other companies will do the same and...

As for evaluation of the performance you need to read the academic papers, Cell is typically 10x (or more) faster than a traditional core - even on problems which appear "Cell hostile".
The only real problem with it is you have to program it specifically to get that level of performance, you can't just take existing code and expect a free speedup.
 

seebs

Newcomer
If I had a LOT of them, I could put them in the garage. One or two, they couldn't keep it warm enough and the cold would kill them through condensation.
 

Nite_Hawk

Veteran
If I had a LOT of them, I could put them in the garage. One or two, they couldn't keep it warm enough and the cold would kill them through condensation.

Yeah, it's pretty cold this weekend. Ready for the snow? :p

Btw, where in Minnesota are you? I'm amazed there's someone else on these boards besides Geo from around here. :p

Nite_Hawk
 

seebs

Newcomer
Northfield, these days. Used to be in Saint Paul.

BTW, are you sure you're from Minnesota? The phrase "ready for the snow" sounds vaguely ungrammatical to me. :)
 

Frank

Certified not a majority
Veteran
Does it have a chance?
It's a great upgrade to the most popular processor in existence: the ARM. And it will run the target apps without much modification, giving you time to put all that added power to good use. It's also a great upgrade for the high end stuff: servers and supercomputers (as said).

Basically anything that uses Linux or similar (BSD etc) would be a great target. But you won't see it replacing x86 on the desktop any time soon, if at all.
 

3dilettante

Legend
Alpha
What exactly does Cell have in common with ARM?

There's practically no code that can run well without modification on Cell, unless you want to run it all on the PPE and waste two thirds of the chip.
The ISA change, threading, and the need to refactor algorithms is not trivial.

Cell is a terrible choice for most of ARM market. It's big, complex, expensive, and probably leaks more current in standby than most ARMs draw under load.
 

Frank

Certified not a majority
Veteran
Definitely.

I meant, that if you have a black box that runs a Linux derivate and you run out of steam with the ARM, a Cell would be a good upgrade. That's a small part of that whole market, but still a serious amount of boxes.

And your application would most likely run after a recompile (although only on the PPU), making the transition a lot easier.
 

Gubbi

Veteran
Definitely.

I meant, that if you have a black box that runs a Linux derivate and you run out of steam with the ARM, a Cell would be a good upgrade. That's a small part of that whole market, but still a serious amount of boxes.

If you start out using an ARM processor, you are worried about:
1. Cost
2. Power usage
3. Performance

Likely in that order. CELL is expensive and power hungry. You'd probably be much better off with a G3/4 PPC derivative or a MIPS core. - Or multiple ARM cores, the new A9 with dual issue OOO execution and +1GHz operating frequency core look promising (and all in 1.5mm^2 in 65nm to boot)

Edit: Alright, just echoing 3dilettantes points

Cheers
 
Top