Initial Core benchmarks look dire for AMD

Initial Core benchmarks look dire for AMD

Summary: In a shootout that tested some of the more popular games and video compression Codecs, Intel's 2.66 GHz Conroe bested AMD's Athlon by an average of more than 20%!

SHARE:
TOPICS: Processors
145

In an AnandTech article yesterday comparing a top of the line AMD dual-core Athlon 64 FX-60 overclocked at 2.8 GHz (2.6 GHz stock) to a sample Intel dual-core Conroe E6700 2.66 GHz processor, AMD's top dog took a devastating beating.  To make matters worse for AMD, this particular 2.66 GHz sample from Intel isn't even the fastest CPU planned for release in Q3 of this year.  In a shootout that tested some of the more popular games and video compression Codecs, Intel's 2.66 GHz Conroe bested AMD's Athlon by an average of more than 20%!

While Intel won't specify exactly how fast the top of the line Conroe processors will run, numbers like 2.8 or even 3.0 GHz have been floating around.  AMD will be making enhancements on their Athlon line by adding memory bandwidth with DDR2 memory as well as lower their processor's power consumption, but nothing as dramatic as Intel's shift to the Core architecture.

These high marks for the Conroe desktop processor can also translate in to great performance numbers for Merom on laptops and Woodcrest on the servers if their clock rates are anywhere close to Conroe.  Mobile and server CPUs are usually clocked slower than their desktop counterparts so we'll have to wait and see how close and what AMD's response will be.

Topic: Processors

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

145 comments
Log in or register to join the discussion
  • AMD will comback by the time this comes out

    I dont think AMD will let themselves be beat. After all the stuff they have been doing. Either way they werent expecting the new Intel chips to be as fast as Intel claimed. But intel just made a mistake by letting the news leak out because now AMD has time to catch up.
    Jimster480
    • AMD may very well catch up some day, but in 6 months?

      "But intel just made a mistake by letting the news leak out because now AMD has time to catch up"

      And you think they'll do this in 6 months?
      george_ou
      • Six months later, INTEL will still be five generations behind

        Whole INTEL is so amateurish compared to AMD:

        http://www.theregister.co.uk/2006/03/10/intel_heat/page2.html

        http://sharikou.blogspot.com/2006/03/intel-fellow-choked-on-question.html
        sharikou
        • You Are out of touch.

          AMD did a real nice job on their recent batch of chips, that took alot of work, dont kid yourself. ALOT of work. Now your saying in six months they will be 5 generations ahead?? Your either kidding, or your just a kid who dosnt know what your talking about.
          Cayble
      • George, its only 20%

        George, you say the AMD overclocked desktop design is 20% behind the yet to be released Intel system.
        That is not the end of the world. AMD has better, faster chips coming. Of course Intel is going to focus on one-socket desktop machines, because their SMP bus architecture is ancient (basically it is a ten year old Pentium Pro bus, albeit overclocked).
        On the server side, expect similar Intel comparisons to one and two socket systems. They will still perform poorly on four socket benchmarks until CSI comes out on Xeon in 2009.
        Also, realize AMD has quad core chips coming in 2007. Intel does not. Sure, they will put two Woodcrests on a single module, and call it quad core, but that is like calling four cars and a motorcycle an 18-wheeler.
        meh1309
        • 20% is enough

          The worst part is, that wasn't even their top end product.

          So what.. My Car is a Super-Charged V6. The Supercharger is only 40 additional horsepower, but 40 HP is enough to notice on a 200 HP engine. That is around 20%. When you are trying to process a huge file that takes 5 hours, you would probably want to spend the extra money to make it process in under 4 hours. Especially when you consider the cost of the individual running the application. Say we are talking a $100 dollar an hour contractor. His time would be saved the cost of a $500 dollar more expensive processor in just 20 hours of work time. That's not even a week to recooperate that cost.

          20% is a big deal.
          nucrash
          • Well, 20% is nice...

            20% is NOT a big deal when intel is not going to release their higher end chipsets 7 months AFTER AMD releases theirs.
            And you really cant compare a car to a PC...they are completely different systems.
            UB3RATL4Sf00
          • 20% is HUGE

            considering that Intel is down about 20-25% now and everybody thinks that AMD vastly outperforms Intel with current generation ...

            I myself am a huge AMD fan and this was a shock to me and I would have had a hard time believing it if not for Anand performing the tests. There's no beating around the bush, in 6 months Intel will regain the performance lead. Still, only an idiot or someone that owns intel stock will be happy about this news because we need a strong AMD to keep pressure on Intel or ELSE ...

            As much as I'd like to believe, I don't think AMD has any chance of keeping up this year. I can only hope that they will recapture the lead next year because let's not forget, Conroe is the equivalent of Athlon 64 and Athlon 64 is already more than 2 years old therefore in the next 18 months AMD will probably launch the next gen. If Athlon 64 is K9 then conroe would be a P5, anything before that was just a tweak for the p4 architecture.

            The problem is that now all AMD partners will go and ask them "Is it true? Will you have anything to counter Conroe in 6 months?" and AMD absolutely must be honest when answering, they can't afford to lie like Intel does because their reputation is just starting to improve ...
            cgrecu
          • Too True

            Well, you have a point. I am a AMD nutjob as well, but Intel has been trying to get their crap together since the older Xeons failed the XP64 test.

            By-the-by, a little trivia: Pentium 5 was the code name for the Pentium Pro, this was before P2 and P3 even, but they skipped ahead, so the series would actually become P6
            UB3RATL4Sf00
          • That's Bass Ackwards!

            [b]Especially when you consider the cost of the individual running the application. Say we are talking a $100 dollar an hour contractor. His time would be saved the cost of a $500 dollar more expensive processor in just 20 hours of work time. That's not even a week to recooperate that cost.[/b]

            Ok... So let me see if I got this right. Let's say you're doing a task for a client. You're charging the client $100 per hour to run a process on his data. In which case, wouldn't you want to make the job take as LONG as possible? If you make it so that the 5 hour job gets done in 4, you've LOST $100, not gained.

            If anything, common sense dictates you should rather the job take 6 hours instead.

            Unless you've got a backlog of similar projects lined up and need to move them out quickly, and are charging a fixed price - not a per hour basis will you be making money by getting a faster, more expensive processor.
            Wolfie2K3
          • Not really

            I'd imagine he's talking about an on-site contractor where the company paying supplies the computer.

            Although really common sense dictates that if one contractor takes 6 hours to do a job, and the next one takes 4 hours to do the same job, which one are you going to hire next time?
            StevoCJ
          • American Line of Thought?

            If I take my time, I can get more of an hourly rate.

            Instead of, If I get my work done, I can finish up with this job, and move on to the next, and get a reputation as a high-quality worker.

            I guess this is why are doing well in the global environment.

            Also, this would be from the employer perspective, not the employee perspective.
            nucrash
        • More specifically 20% better than FX-60

          Yep. Intel's somewhat suspicious set of tests appear to show that Intel's nextgen offering is 20% faster than existing AMD chip.

          Given that this represents something like a 50% improvement in Intel's performance compared to what it had last year, why would it be unreasonable to expect that in another 6 months AMD will release AM2 socket, DDR2 memory support and a variety of other upgrades (which they have been holding back because they didn't NEED them to be beating Intel) and see at least a 20% increase in their own speed?
          spark555
  • All the tests were timedemos

    None of the tests were done with real world applications. When the Conroes hit the web review circuit, they will be tested in real game environments and monitored with FRAPS. Then the true performance rating will be achieved. Also remember that AMD had the technology lead and is now refining it further. Intel had to do a total about face. Woodcrest/Conroe/Merom is their true multicore solution. All previous offerings were stopgaps to curb loss of market share.
    etrigan63
    • What do you think timedemos are?

      Timedemos are as real as it gets. Also note the video encoding times for Apple and Microsoft Codecs. These numbers are all real, and you'd be foolish to think that AMD execs aren't sweating now.
      george_ou
      • Intel simply lacks credibility

        Given the fact INTEL is losing market share at alarming rate, it's quite desperate to win back customers' money.

        Read the following:

        http://sharikou.blogspot.com/2006/02/beware-of-intels-tricks.html

        http://www.vanshardware.com/reviews/2002/08/020822_AthlonXP2600/020822_AthlonXP2600.htm

        http://www.swallowtail.org/naughty-intel.html

        http://blogs.zdnet.com/ip-telephony/?p=947


        http://sharikou.blogspot.com/2006/03/post-idf-news-summary.html
        sharikou
        • You just fail to admit defeat

          AMD can lose another battle. That wouldn't be the end of the world for them. They atleast took top honors for once. As opposed the the Linux Desktop world who just continually gets pulverized by Microsoft. Or Macintosh for their recent security problems.

          AMD has proven that they can build a faster processor. Let Intel take back their moment of glory and watch AMD come back with something even better. May not be for a couple years, but AMD has some brains in R&D. They will prove this again with the first release of Quad Cores.
          nucrash
          • Feed The Need

            Your exactly right, this is what the processor market NEEDs, if Intel gives AMD something to fight for, Im pretty sure the overall results for the consumer will be good.
            UB3RATL4Sf00
          • Great post

            You're exactly right. The CPU market is due for some competition. If this forces AMD to be better, then we the consumers benefit. Intel has been losing the CPU wars for the last 2 years. It's just their turn to shine for a while.
            george_ou
          • What the...

            This thing about you and I agreeing has to stop...

            How am I suppose to carry on mindlessly if your points are valid.
            nucrash