Core Duo Reaches the Desktop

rtt writes "AMD has long reigned the desktop CPU market due to Intel's offerings struggling to keep up in terms of performance and power consumption. Yonah is the predecessor to the Core architecture and is predominantly a mobile chip, and is used at the heart of Intel's Viiv technology. Bit-tech has an article about Yonah beating the top of the range desktop AMD chip, the FX60, clock for clock. From the article" 'When Yonah is running at the same clock speed as AMD's Athlon 64 FX-60, we found that it beat it into a corner in just about every situation.'"

Comments

every situation?

[Anonymous Coward]

mp3 encoding and other floating point workloads are quite common. AMD wins hands down there.

Wake me up when it supports 64-bit

[RelliK]

When it supports AMD64 instruction set, it will be worth a look. Until then who cares?

Re:Wake me up when it supports 64-bit

[myurr]

Not only is this chip 32bit only, it's also dual core as opposed to the FX60's single core, its been built on a 65nm process unlike the FX60's 90nm, and the FX60 is actually starting to show it's age a little bit. I big giant "Duh" springs to mind at this point.

It's a play on words.

[insomniac8400]

"When Yonah is running at the same clock speed as AMD's Athlon 64 FX-60, we found that it beat it into a corner in just about every situation." If this is true, it would be the first time intel made anything better than amd. But in the end, all that matters is that AMD's $200 chip outperforms intel's $200 chip.

Re:It's a play on words.

[anagama]

It's worth noting that the comparison is between the FX-60 running at factory speed, and a Core Duo running overclocked.

Re:It's a play on words.

[spleck]

It's worth nothing that the comparison is between the FX-60 running at 2.6 GHz, and a Core Duo running at 2.6 GHz. Hence the clock-for-clock comparison. I think they were trying to compare architectures, for which I thought the article did a good job. I learned that the Yonah is nice, but can't do 64-bit or FPU operations well.

Actually, I already knew that, but I still looked at the benchmarks.

Over the top

[Rorian]

I think they go just a bit overboard in saying that Yonah beats the FX-60 "into a corner".. Most benchmarks had it either infront or behind by around 2-3%. Is it really worth forking out a few hundred dollars for such a dismal gain in performance? Does it have better performance-per-watt? That's what really seems to count these days anyway.

Re:Over the top

[jsoderba]

Is it really worth forking out a few hundred dollars for such a dismal gain in performance?

The Core Duo is expensive, but have you looked at what AMD is charging for the FX series? Intel wins on price/performance by a mile. On the other hand, they're overclocking the Intel CPU a hell of a lot and running the AMD CPU at stock speed, so it's not exactly a fair comparison. Still says a lot about what we can expect from Core 2, though.

Does it have better performance-per-watt?

What? We are talking about the Core Duo here, the most efficent x86 CPU on the market.

Re:Over the top

[Aadain2001]

The overclocked the Core chip to MATCH the clock frequency of the FX60 chip. It's not an unfair comparison at all. They eliminated the clock as a difference and instead are only comparing the performance of the architectures, clock-per-clock. Viewed like that, the Core chip is better (not by a huge amount, but still better). Through in the lower power & heat and lower price, and the Core chip beats the FX60 by a very wide margin.

Dubious Test

[cait56]

As near as I can determine from reading the article, it proves that a Core Duo *slightly* outperforms an Athlon 64 XP2 when doing heavy number crunching with a 32-bit Windows application.

Comparing the same application build for 64-bit on Linux vs. 32-bit on Linux (or BSD) would have been a far more meaningful comparison.

Until Conroe

[SuperKendall]

Yes, until Conroe arrives later this year.

Until then the consumer space really doesn't need 64-bit processors for most work people do.

Just shows how selective statistics can lie

[jmorris42]

Ok, Intel is getting some game, 'bout freaking time. But this isn't an "AMD killer" by any means.

Notice how they only included ONE FPU intensive task and AMD (and several of the Intel products) schooled this mobile offering? Most reviews include a lot more balanced set of tests, this one obviously had their storyline written for them and was tailoring the tests to fit the plot.

And also, let us not forget that the STOCK benchmark numbers for this chip were anything but impressive, so they played up the overclocked numbers. However, while this chip does have some seriously intersting overclock potential it isn't the first chip to be massively overclocked. Just last week we were salivating over a budgie Intel chip that somebody overclocked into the world's fastest CPU. So why not include THAT firebreathing monster's numbers on the chart along with some seriously overclocked AMD parts? Perhaps that would't have had such a dramatic narrative? Ah.

Meanwhile, I'll keep comparing parts running at factory spec and waiting to see what AMD drops next week to compare current gen parts to current gen parts.

hrm

[Silicon Mike]

So.. What they are saying is Intel's latest and greatest CPU is finally gonna beat something AMD released, what, 5 or 6 months ago? Intel is STILL playing catch-up.

Enter obligatory comment

[Temujin_12]

Entry obligatory AMD zealot, "intel can't possibly make something better then AMD" comments here _______.

Seriously, I'm sick of the overly zealous statements when it comes to "OS vs. OS" "Company vs. Company" etc. debates. Why is it so threatening when another companies/organizations happen to produce something better than your favorite company/organization?

Sometimes intel will get it right and sometimes AMD will. Deal with it.

Ya, I know. This is probably eligible for flaimbait and/or troll mod points. Oh well, I just needed to get this off my chest.

Re:What the heck is "Intel's ViiV technology?"

[mrchaotica]

Unless that means it doesn't support DRM?

On the contrary, it means that it does support DRM. And not just DRM, but Treacherous Computing. In hardware.

That marketing blurb should read:

Intel says: With Intel Viiv technology, we control a highly locked-down Intel platform designed for protecting publishers' "Intellectual Property". That means you can: Let us take charge of "your" media. Share experiences with movies, photos, and music with your friends and family if we let you. Complicate your digital life.

It's 1984-style DoubleSpeak.

Re:Sham

[Clockwurk]

As others have pointed out, the Core Duo only beat out the Athlon64 FX-60 when overclocked.

Both were clocked to 2.6ghz so that a clock for clock comparison could be made...

Re:Sham

[Malor]

They're misrepresenting the product. I have to wonder if they were paid for this review.

Of course they were.... If not outright in cash, by getting early access to hot new stuff.

It's also not unknown for a manufacturer to 'accidentally' forget to ask for their stuff back if you write a really glowing review of it.

NVidia appears to have shills working the forums, hired via some marketing agency. This is a hard thing to prove conclusively, but there was at least one documented case a couple months back, so assuming that there are more seems reasonable. This could be just a new evolution in that process... shill websites.

It's interesting that ALL of the preview articles I've read have involved massively overclocking the Conroe, and then breathing hard about the OCed numbers. ALL of them. Nobody seems to talk about stock speeds much at all.

Re:Idiots of the world unite!

[Anonymous Freak]

While it is true that for x86 processors, 64-bit mode is measurably faster than 32-bit mode, it's not MASSIVELY different. (i.e. The Athlon FX-60 is bested by the Core Duo T2500, but not by much, running the AMD in 64-bit mode with native 64-bit apps, the AMD would probably barely beat the Core Duo then. The main speed increase doesn't come specifically from being 64-bit, but rather that the 64-bit extensions more than double the number of registers. THAT is what makes it faster. If AMD had added the registers into a 32-bit mode, the 32-bit mode would be just as fast as the 64-bit mode.

So I do partially agree with the OP, but he does seem to be putting too much stock on 64-bit mode. It's not like the moment Vista comes out, all 32-bit processors will be obsolete immediately. Companies will continue to produce 32-bit apps for some time. Heck, it took nigh on 3 years after the intro of OS X for companies to stop producing OS 9 versions of their apps, and that is a significantly harder changeover! Microsoft is famous for eternal compatibility, so they won't be pushing for 64-bit-only apps for some time. (How many Windows users are STILL running on Pentium 3s with Windows 98?)

To use an even more recent example, Apple has switched to Intel processors (In all likelihood, the entire line will be switched over by July.) Yet some MAJOR applications (Microsoft Office and the Adobe apps,) will likely not be Intel-native for at least another year. The change from 32-bit to 64-bit in the Windows world will produce about a 10-15% increase in speed; yet the improvement in the Mac world can produce up to a 400% increase! You'd think Adobe would have had their ass in gear! I don't imagine that MOST Windows developers will switch to 64-bit until the majority of the Windows computers (both brand new and 'already in service',) are 64-bit. (Then again, 64-bit Windows has been around for over a year now, and how many have 64-bit apps?)

Reign on Your Parade

[Doc Ruby]

"AMD has long reigned the desktop CPU market"

AMD has never (or infrequently and briefly) ruled the market. Their often superior tech might sometimes "lead the pack", or even "lead the market" in the sense of directing development. But more people buy Intel desktop CPUs, which is what rules the market.