Intel Conroe vs Cell

Wollan

Member
I personally don't have much clue on how to do a comparison but Im sure there are several here at NeoGAF willing to give it a shot. Maybe this is a ridiculous comparison(Cell mostly being geared for graphical/physics operations for example), I don't know, of course there are issues like the Cell most likely being harder to harness power from(thanks to Conroes single/dual core nature) but it would be fun to hear what people think.

Im specifically talking about the PS3 Cell(1x PPU, 7x SPUs @ 3.2ghz).

After looking at several Conroe benchmarks online, it seems to best it's AMD equilants(in mhz atleast) at around 25-30%(very generally speaking, there are a ton of different benchmarks).

Would be nice if anyone could try to give a comparison, theory or thoughts between the two cpu's for game-applications.

000163281.jpg


CELL_2.jpg
 
thing is cell is old news now really....don't expect anything other than ps3 using it. Just another over-hyped piece of development from sony.





{troll}

actually I pretty much believe that cell isn't all its cracked up to be.
 
Seriously this belongs at Beyond 3d...cause gofreak aside there really aren't any technical guru's on this forums..

All you will get is is idiotic responses like "OMG CELL = EMOTION ENGINE 2" "SONY AM ROSEBUDDED" etc...

and a bunch of Xbots claming that this chip is superior to cell even though they don't know what they are talking about. Therefore if you want to have a serious dicussion please take it to BEYOND 3d :D
 
Comparing two different architectures (x86/PPC) is dumb enough, what's even worse is comparing two completely different design philosophies. Personally, i favour Cell's multicore design over what's available on the PC right now, but what effect that could/would have on performance is pure speculation.
 
Both fail to work with my Socket 939 motherboard so I'm pretty sure they are crap. If they are actually very good (like say... the Conroe offering amazing price and performance that X2's cannot compete with even after the current price drop) I won't know because my I'll be in the corner with my scant upgrade budget stuffed in my ears.
 
I would think the Intel chip would be easier for developers to work with than either the Xbox 360 or PS3 CPU solutions. Hard to imagine the chip couldn't do just as well if not better than either of the console CPU's :).
 
_leech_ said:
Comparing two different architectures (x86/PPC) is dumb enough, what's even worse is comparing two completely different design philosophies.

you have to compare them by a common metric, so the obvious answer is weight. the lighter one is a witch.
 
Well as you can see, the Cell comes in all kinds of weird psychedelic colors and the Conroe is very plain and boring, therefore it wins
 
Doom_Bringer said:
oh you work for Intel?? Sorry! Didn't mean to diss your company
I'm not that sensitive about it, really. But (apparantly) Conroe kicks ass. Intel has been amazingly dumb, and has had inferior products, for a while now. But not for the next year or so going forward. :)
 
It's not comparable until there is an application that runs on both architectures, and unless the Mac switches to Cell for some reason or Linux is ported over, chances are slim

However, it's doubtful IMO that the Cell will hold any kind of lead except in marketability as far as pure horsepower and flexibility goes. There might be certain multithreaded-heavy apps with a smaller memory footprint that run much better on Cell though (Like physics)
 
Wollan said:
(Cell mostly being geared for graphical/physics operations for example)

Actually the Cell is NOT all that great at physics because of the incredibly limited cache (256k per SPE and 512k for PPE... I think). Either way, the Cell is a specialized processor whereas the Core2Duo is a beast at general processing... Plus the Core2Duo is an Out-Of-Order processor while the Cell is an In-Order processor (instantly cutting it's realworld performance by at least 25% -- that's what Carmack says anyway). Then of course there is the whole hassle of programming for the Cell, which would be more difficult than programming for the Core2Duo, meaning it's easier to be more efficient with the Core2Duo's resources (it has 4megs of cache, almost double that of the entire PS3 Cell) squeezing out even more performance.

I'd be willing to bet that even if the theoretical numbers for the Cell are two, three, or even four times that of the Core2Duo, in real-world performance the Core2Duo would be faster by a sizeable margin.

GhaleonEB said:
50% through the Intel Store.

Those Core2Duos are looking mighty tempting - at %50 off it'd be a must buy. Care to do me a favour?
 
White Man said:
Can you comment on the rumor of quad socket (8 core) Woodcrest configurations?


http://hardware.slashdot.org/article.pl?sid=06/07/27/1753255&from=rss

"Core scales and it will be scaling to the level we expect it to. That also applies to the upcoming generations - they all will come with the right scaling factors. But, of course, I would be lying if I said that it scales from here to eternity. In general, I believe that we will be able to do very well against what AMD will be able to do. I want everybody to go from a frequency world to a number-of-cores-world. But especially in the client space, we have to be very careful with overloading the market with a number of cores and see what is useful."
 
Just to point out ... CELL was always designed to be a scaleable architecture. The PS3 CELL is obviously not the high-water mark for the design.
 
Onix said:
Just to point out ... CELL was always designed to be a scaleable architecture. The PS3 CELL is obviously not the high-water mark for the design.
But do you think it could AFFORDABLY in the next, say, 5 years have a design that holds a lead over the upcoming AMD/Intel offerings, even if they continued to scale up?
 
Doom_Bringer said:
Seriously this belongs at Beyond 3d...cause gofreak aside there really aren't any technical guru's on this forums..

To be fair, there are a few others on this forum that know what they're talking about when it comes to tech. Mmmkay and Panajev come to mind.
 
Seriously Doom_Bringer check out the benchmarks. Not only is Conroe faster, but it runs cooler and uses less energy to power than the AMD alternative (Athlon 64 X2 AM2).

Disliking a product for no real reason (unless you want to give one) is pretty foolish in my opinion. Intel has indeed been sucking for the last couple years, but the Conroe blows that all away. When I upgrade in the near future it will be Conroe and I was an AMD/ATI guy without question before the benchmarks.
 
MurdrOfCrows said:
When I upgrade in the near future it will be Conroe and I was an AMD/ATI guy without question before the benchmarks.

You'll be saying the same thing about whatever AMD chip is released in the 2nd quarter of next year. Seriously.
 
The Abominable Snowman said:
But do you think it could AFFORDABLY in the next, say, 5 years have a design that holds a lead over the upcoming AMD/Intel offerings, even if they continued to scale up?

I'm not sure what IBM's intent is for future generations. The current CELL at least, is not really meant for true general purpose use - so directly scaling its to the Nth degree probably wouldn't make sense for that usage (long term).

Until we see what IBM has in store, I'm not sure the answer is evident.
 
I really thought Cell wasn't going to turn out to be the Emotion Engine 2. I think even IBM has no plans to use the thing in the future. :P

Better luck next time Kutaragi, if there is a next time for you.
 
In the end, Cell is definitly the bigger potential due to the number of SPUs but if those SPU's had the flexibility of the Conroe cores, the Cell would be a whole different beast.
 
-Rogue5- said:
Actually the Cell is NOT all that great at physics because of the incredibly limited cache (256k per SPE and 512k for PPE... I think). Either way, the Cell is a specialized processor whereas the Core2Duo is a beast at general processing... Plus the Core2Duo is an Out-Of-Order processor while the Cell is an In-Order processor (instantly cutting it's realworld performance by at least 25% -- that's what Carmack says anyway). Then of course there is the whole hassle of programming for the Cell, which would be more difficult than programming for the Core2Duo, meaning it's easier to be more efficient with the Core2Duo's resources (it has 4megs of cache, almost double that of the entire PS3 Cell) squeezing out even more performance.

I'd be willing to bet that even if the theoretical numbers for the Cell are two, three, or even four times that of the Core2Duo, in real-world performance the Core2Duo would be faster by a sizeable margin.

Everything you just wrote is complete bullshit.

There wouldn't be a company selling a Cell like add-on board for x86 PCs if anything you wrote was remotely true.
 
GhaleonEB said:
I'm not that sensitive about it, really. But (apparantly) Conroe kicks ass. Intel has been amazingly dumb, and has had inferior products, for a while now. But not for the next year or so going forward. :)

Notwithstanding Conroe, why has Intel had so many goof-ups the past few years? It seems to me that it all started with the P4 marchitecture. Here's my theory of what REALLY happened:

Intel marketing:
"All right listen up you nerdy, slide rule toting dweebs, our market research shows that raw GHz sells."

Intel Engineering:
"But, just because you spin the engine faster doesn't mean you get as much work accomplished clock for clock." And then you have thermal efficiency problems... You know, H E A T.

Intel marketing:
*Dazed look* "Moving right along then." "So we're gonna' build this new processor, and it's REALLY fast."

Intel Engineering:
But what about current leakage?

Intel marketing: Just Fast!
Intel engineering: But...
Intel marketing: Sh!
Intel engineers: I'm...
Intel marketing: Sh!
Intel engineers: All I'm say...
Intel marketing: Sh!
Intel engineers: There gonna get a...
Intel marketing: Sh!
Intel engineers: We’re just...
Intel marketing: Sh!
Intel engineers: Would...
Intel Marketing: Sh!... Knock-knock.
Intel engineers: Who's there?
Intel marketing: Sh!
Intel engineers: But...
Intel markeing: Let me tell you a little story about a man named Sh! Sh! even before you start. That was a pre-emptive "sh!" Now, I have a whole bag of "sh!" with your name on it. :)
 
jett said:
I really thought Cell wasn't going to turn out to be the Emotion Engine 2. I think even IBM has no plans to use the thing in the future. :P

Better luck next time Kutaragi, if there is a next time for you.

http://www-03.ibm.com/press/us/en/photo/19199.wss

"A nine-core IBM Cell blade

The first Cell-based product from IBM, designed for businesses that need the breakthrough multi-core architecture capabilities of the IBM Cell Broadband Engine processor to tackle tasks involving compute-intensive workloads and broadband media applications."

http://news.com.com/IBM+signs+outside+customer+for+Cell/2100-1006_3-5765212.html

http://www.techspot.com/news/20394-ibm-has-potential-cell-customers-in-defense-contracting.html

http://www.research.ibm.com/cell/

http://www-128.ibm.com/developerworks/power/cell/downloads.html

And that is just from a minute of googling around IBM's website. There is an article from earlier this year that talks about IBM having over 200 companies they are working with to provide Cell chips for.
 
Marathon said:
Everything you just wrote is complete bullshit.

I agree with you on that...

There wouldn't be a company selling a Cell like add-on board for x86 PCs if anything you wrote was remotely true.

...but I'm not sure if that really proves it.
 
Marathon said:
http://www-03.ibm.com/press/us/en/photo/19199.wss

"A nine-core IBM Cell blade

The first Cell-based product from IBM, designed for businesses that need the breakthrough multi-core architecture capabilities of the IBM Cell Broadband Engine processor to tackle tasks involving compute-intensive workloads and broadband media applications."

http://news.com.com/IBM+signs+outside+customer+for+Cell/2100-1006_3-5765212.html

http://www.techspot.com/news/20394-ibm-has-potential-cell-customers-in-defense-contracting.html

http://www.research.ibm.com/cell/

http://www-128.ibm.com/developerworks/power/cell/downloads.html

And that is just from a minute of googling around IBM's website. There is an article from earlier this year that talks about IBM having over 200 companies they are working with to provide Cell chips for.

The Blade server, the defense "potential" and the sensors or whatever, that's all old news.
 
Top Bottom