AnandTech has published part II (part I discussed here on OSNews) of their performance preview of Intel’s upcoming Yonah. “We’ve updated the benchmark suite considerably, including modern day games and a few professional-level applications hopefully to get a better perspective on Yonah’s performance. We’ve also included an Athlon 64 X2 running at 2.0GHz, but with each core having a full 1MB L2 cache, making the Yonah vs. X2 comparison as close to even as possible (not mentioning the fact that AMD has twice the advantage in this round, with both a larger L1 cache and an on-die memory controller, but it should make things interesting).”
Given how Intel has basically had their head wedged up their backsides on instruction optimization for the past five years, I am genuinely shocked to see the 2ghz Yonah perform as well as it did against the A64. While it was still edged out, that intel has FINALLY made serious execution optimizations shows that maybe… just maybe the giant has been roused from it’s slumber.
I flat out would never have expected it to edge past the Pentium M in any benchmarks, much less challenge the X2 3800+. From those benchmarks I would venture a guess at it being a much closer match speed-wise to the 1.82ghz “Venice” A64 3000+
I flat out would never have expected it to edge past the Pentium M in any benchmarks, much less challenge the X2 3800+. From those benchmarks I would venture a guess at it being a much closer match speed-wise to the 1.82ghz “Venice” A64 3000+
You’re probably the only one who has drawn those conclusions. Is this the first time you heard about the Yonah? Or do you have trouble understanding CPU architecture and design? You see, Yonah is a dual-core chip. Both cores run faster than the A64 3000+, have more L2 cache, and can easily keep up with an A64 clock-for-clock.
The Penium-M can hold its own against any AMD64 chip at the same clock rate. Why would we expect less from a dual-core implementation of the Pentium-M? Its not NetBurst, y’know.
If they integrated a memory controller on board I bet AMD would have some serious competition. As it is right now I doubt Intel can offer better price/performance, but probably better power/performance or heat/performance numbers which will be interesting to compare against AMD’s mobile chips, none of which are dual-core yet.
These benchmarks don’t really show the advantages of the Yonah, just the lack of disadvantages for Intel to use this core as a replacement for their failed NetBurst experiment. Whomever decided NetBurst and Itanium was a good roadmap for Intel to take back in 2000 should be sacked, along with anyone that agreed with them.
The future of general purpose personal computing looks cool, quiet and fast! Thanks to frequency-scaling and multi-core CPUs.
or did this comparing a mobile processor against a desktop processor? A mobile processor thats built to run stingy and fast, vs a desktop proc thats built for a$$ blazing fast speed. and folks are talking about amd rocks? wtf? i’ll compare my G35 to your Chevy Aveo and you can brag how much pep the aveo has. Instead of having a boner to be the 1st to break a story, why not have some substance 1st? then maybe it would be a worthwhile read.
To be fair, they are also comparing a top-end $650 processor that won’t be out till next month to a low-end $300 processor that came out months ago. The reason they do this is because Yonah represents the future of Intel’s desktop processors, and people want to see how it stacks up to AMD’s architecture.
I was very impresed with the benchmarks for Yonah. This is Intel’s new, lowend, CPU for laptops. I can’t wait to see how Crusoe and the other new CPUs they have coming stack up against AMD. If Apple does decide to go with Yonah CPUs for the new iBooks (hopefully, soon to be announced at Macworld!) I’ll order one immediately.
No, this is *not* Intel’s lowend CPU for laptops. The benchmarked Yonah chip is a $500 number and will be found in notebooks $1500 and up. *Conroe* is Intel’s next-generation desktop chip, and Merom is Intel’s next-generation laptop chip. These will not be “high-end” versions of Yonah, but rather, newer replacements using a new design.
AMD more or less bet their company on 64bit computing awhile back. They made a good bet and are clearly the leader. Now that intel is starting to catch up, it will be interesting to see what AMD retorts with.
The Opteron blew everything out of the water when it was released, what does AMD have hidden away in their R&D labs? Maybe an 8 core chip to compete with Sun’s new processors?
s/8 core chip to compete with Sun’s new processors/8 core chip to compete to be Sun’s new processors/g
Are gonna CREAM pokey Dell XP boxes. Yea!
But are they going to CREAM pokey Yoneh Dell laptops?
the Pentium M actually beats Yonah on the winstone tests, so basically IDEs and other integer based apps won’t get a boost with this first generation. I don’t do multimedia creation, and all the games depend mostly on the GPU anyway.
I’ll catch the multicores a couple generations down the road.
Why do people continue to refer to Intel’s mobile processor strategy in a manner as to suggest that it has some implied self-restrained inferiority to AMD’s single architecture strategy for its market? Clock-for-clock Yonah completely spanks Intel’s “desktop” line of processors at integer performance and power consumption. It will be the largest intellectual ancestor of Core when Merom ships. This is not a castrated processor.
Browser: Links (2.1pre11; Linux 2.6.10-gentoo-r6 i686; 80×40)
Intel Yonah @ 65nm, and *just* matching AMD X2 @ 90nm? Can’t wait to see what AMD does at 65nm.
Maybe keep up with a Merom?
I prefer AMD myself, but I have to admit Intel’s Pentium-M architecture has potential.
Sorry to ask, but why we need smp laptop? (of course the cpu’s maker dream is to sell something new to drop the old ones and make profit) If the second core is on, of course, the battery life will be sorter (almost half if we have a bad(not high-end GFX). But why we need dual core laptop?
I have a dual opteron workstation so I know the advantage , but for a laptop with 12″ or 15″ screen (IMHO the 17″ notebook is not a “mobile” computer)? My dream laptop would be working at least 12 hours on battery, with dvd playing or browsing internet (wireless). If I need cpu performance I buy workstation.
Since you asked, I find my 17″ notebook (9.5#) a very reasonable trade-off to lugging around a shuttle PC w/17″ LCD, keyboard, etc. when on the road for a few weeks…plugging it in due to poor battery life is a small price to pay. So I am of the opposite view, why does my notebook have to have less power than my desktop. I would stop using a desktop totally if the performance was equivalent, but it never is (graphics, HDD speed, etc. always makes it slower to an equivalent desktop).
Very impressive for a notebook cpu, I am curious as to what AMD will do to counter this. On the desktop, AMD clearly spanks Intel, however in the notebook market they are barely even a competitor to Intel. When AMD comes out with a low power high speed notebook cpu they will have Intel shaking in their boots.
One can only hope nobody ever wins this cpu war and intense competition keeps both parties (AMD and Intel {maybe via?}) cranking their money into r&d and hence building a better, faster, and more efficient product for years to come.
I’m an AMD man myself, but I have to give Intel props on this one, for a notebook cpu, this thing is amazing.
I love competition.
I am sitting in front of a dual G5 Mac and I can tell that two CPUs (or two cores) make a big difference for desktop computing if the OS and the libraries are multi threaded / thread safe.
Mac OS X performs very nice on multi cpu machines, so a dual core laptop will give a huge performance boost!
The Yonah has a 2MB shared L2 cache. That is how it manages to keep up with an AMD64 X2 with it’s built-in memory controller. For single threaded apps (i.e., 99.99999% of all apps), the entire 2MB of L2 cache in the Yonah is available for use. The PM also had 2MB of L2 cache. This compares to the 1MB or 512KB of L2 cache in the AMD64 X2s they tested with.
Intel has a history of throwing more cache at the problem when they need better performance, especially on synthetic benchmarks. The Itanium is a good example of that principle, having up to 9MB of cache on some models.