Turion 64, take two

Earlier this week, we published an article comparing the Pentium M 760 to the Turion 64 ML-44. We have since discovered two problems with the tests, and we’ve revised the article to fix these issues.

First, due to an unusual confluence of events, the picCOLOR and WorldBench tests on the Pentium M 760 were inadvertently run without SpeedStep power management enabled. These tests have been rerun and the graphs updated with the new scores for the Pentium M. That means the revised article has new results for each of the WorldBench component tests, whether there has been significant change or not. The updates change the margin of victory in some cases, and in one case (the Worldbench Photoshop benchmark) the winning and losing positions swap between the two CPUs. Perhaps most notably, the WorldBench overall score moves from a tie to a three-point win for the Turion 64 ML-44. We have modified our benchmark commentary and the article’s conclusions based on these new results.

To ensure the accuracy of the other benchmark scores, we ran all of our benchmarks again on the Pentium M 760. No further abnormalities were found.

Second, it was discovered that the MSI motherboard used to test the Turion 64 ML-44 incorrectly configures the upper voltage range for that processor at 1.5V instead of 1.35V. As a result, the “load” power consumption numbers for the Turion 64 were five watts higher than necessary. We’ve updated that graph, as well.

We originally discovered these issues approximately 24 hours ago, but we wanted to make sure we had everything corrected and our facts straight before publishing an update. Our efforts were somewhat hampered by the fact that AMD has literally no technical documentation on its Turion 64 processors on its web site at the time of this writing, making it impossible for us to verify the maximum voltage of the ML-44. We wound up having to contact an AMD PR rep, who got back to us with the official word. Once we had the official number, we used RMClock to artificially lower the maximum voltage to 1.35V for testing, so we wouldn’t have to wait on an updated BIOS from MSI.

We strive for accuracy here at TR, and when our efforts in that area fail to meet our standards, we strive to correct our mistakes as quickly and completely as possible.

Comments closed
    • vanhoa
    • 14 years ago

    I read this at some website – your comments please
    ———————-

    One of my friends, a stock analyst, was working on a Microsoft (MS) Excel* program to analyze some stock indicators. She had an Intel Pentium® III laptop with about 512MB RAM. She was able to run the program for a few stocks, but it stalled whenever she tried to run it with more than 70 stock tickers. She asked me for help in looking into the problem. I tried to run the same program on my personal laptop which is a Pentium M 1.6 GHz with 512 MB RAM. Surprisingly the program ran and completed in two minutes. Based on these results, she decided to upgrade to a newer laptop.

    She went to a major retail chain store and she was flooded with sales people pushing her to buy an AMD-based laptop, She was given all kinds of reasons—the AMD chip is more powerful than any Intel processor out there, and the upcoming Windows Vista OS* which will be a 64-bit OS, and the x64 AMD chip would make an easy transition to the new OS.

    She took the bait and did buy an AMD Turion* with 1GB RAM. Now the fun started. She loaded the machine with MS Office and started running the Excel macro. To her surprise the macro never passed the fifteenth stock ticker even after 30 minutes of running. In pretty much a frustrated state, she called me up and asked my help in looking in to it.

    I tried to compare the machine stats such as process utilization, memory usage, and memory paging side by side on my Pentium M laptop and her AMD Turion. Both were maxing out the processor utilization and both were using about 300MB of memory, but my Pentium M had 720 threads running and her AMD Turion had 310 threads. We were really curious, because supposedly the Turion outperformed the Pentium M. Finally we decided to take both our laptops back to the chain store and get them checked out by the techies out there.

    At the store, my friend called to the floor manager in the laptop section. We explained what we were seeing with the program on our laptops. The techies started to explain to us that AMD processors are way better than Intel’s and that they beat Intel processors in every aspect. The manager pulled in a customer who was vouching that he would not buy anything but AMD because they were such awesome machines. So we opened both our laptops and started the Excel macro on our laptops simultaneously. Soon a small crowd gathered around us to see what was going on.

    My Intel Pentium pretty much crushed the AMD Turion in running the macro. The techie was in disbelief. He started saying that there might be something else running on the AMD machine slowing it down, so he uninstalled the virus scan and removed everything that could be safely removed. We then did the same test and the results were the same—my Intel Pentium laptop out performed the AMD Turion by at least 10 to 1. The people gathered around were really surprised, as all of them were being sold AMD products by these same techies who weren’t able to explain why the latest AMD Turion was slower than my 8-month-old Pentium M 1.6 GHZ processor.

    Finally another guy came out from the help desk with a probable answer. He said that Microsoft Office was optimized for Intel chips to take advantage of the way Intel chips handle floating point logic, and that AMD was faster in processing speed either way. Before I could ask, a person from the crowd who said he is a business user who uses Excel often, asked them why he would want to buy an AMD for sheer processing power when it would not perform well for the applications that he wants to use his computer for. They didn’t have a good answer for that.

    I think I was able to help at least 10 to 15 people in that crowd decide what to buy. We depend on our computers to run applications that are vital for our daily business. I think these stores are leading buyers to buy AMD, and I don’t know if they are educated enough on what the key differences are or are they aware of the needs of their customers and the products that would serve them well.

      • Gaidheal
      • 13 years ago

      Nothing in life is black and white. I vastly prefer AMD to Intel on a whole range of issues but when buying any tool you need to make sure it has the features you require and that you are not paying a premium for features you will not use.

      With respect to your story, there is no way it was a 10:1 ratio at all, however it could very easily, based on the differing technical details and assuming optimization for Intel operations, be 2:1 which is significant. Bear in mind however, that floating point maths is *not* actually used very much by most business users at all. Typically they are word-processing, browsing and running various other processor intensive, *not* FP intensive, applications. Unless your expected usage is going to rely heavily or exclusively on FP calculations there is no good reason to buy an Intel over an AMD at the same price point; you can pretty much always get a better overall performance-per-penny deal with an AMD based system.

    • obajemu
    • 14 years ago

    Thanks for making me see the light about Turion

    • Hector
    • 14 years ago

    Thanks for taking the time to retest. Only other change i’d still like to see is memory timings ajusted 760 @ 3-2-2 and Turion @ 2-2-2 1T which that ballistix can do in both cases. A64 loves low ltency.

    • WaltC
    • 14 years ago

    Good job all around. It isn’t as widely recognized as it should be, but in-depth comparisons like this which cover a gamut of complex hardware and software component interactions are full of details that often aren’t immediately obvious. The topics covered in these articles are indeed complex and deserve the thorough treatment you’ve given them. In cases like these, especially–the devil’s in the details, no doubt about it. Thanks for the efforts to keep us informed!

    • mark5004
    • 14 years ago

    Compare like for like.

    The Turion is 2.4 and you could have tested a 2.26Ghz 780 or you could have overclocked the chip to 2.4Ghz.

    You would have not liked the results because the Turion would have been crushed.

      • Hattig
      • 14 years ago

      Um, I thought we had already got over MHz as the sole metric for reviews years ago?

      Quite clearly price is a suitable metric to use in this situation, and certainly applies to people looking at the two CPUs for SFF or media systems. In this case the processors might be priced quite closely, but the platforms aren’t. Another metric is power draw / battery life.

      In terms of power draw, I expect that things will get closer when Turion gets to use DDR2, which can use up to 7W less than DDR1 (DDR2-666 vs DDR-400). Blame AMD for not DDR2-ising Turion sooner.

        • wierdo
        • 14 years ago

        lotsa people are still stuck in the stone age of cpu speed measurements.

          • tfp
          • 14 years ago

          When they are two different architectures one not being reviewed here before it would be nice to see how they perform on a clock-by-clock comparison. That way it is easier to estimate how the chips will perform considering we have no model numbers that can really be used as a guide.

          At the worst it would have been nice to see how the best of each compared just like they do with every other processor benchmark. It would have also been interesting to see how different clock speeds of the Turon and P-M compare in power consumption.

      • Hector
      • 14 years ago

      y[http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2648<]§ As you can see 2.0ghz X2 beats 760. But it's a stupid comparision anyway - Should we have set processors to same clock speed for years with the p4 vs. athlon xp and 64 comparos? I don't think you'd like that either.

        • tfp
        • 14 years ago

        I thought people did at one point just to show the difference. But then compared the fastest of each type to show what they can do. An architecture comparison and a processor comparison.

    • dgage
    • 14 years ago

    As usual, great work with integrity. Thanks!

    Now, what I would really like to see is the best Turion against a similarly priced (or even more expensive) Duo Core chip to compare the latest offerings. Would definitely be interested in the performance running multiple applications simultaneously as we are all apt to do. Any chances of having that done in the next month before we need to buy more laptops? Please, please!

    • Stefan
    • 14 years ago

    This makes the Turion look even sweeter. I would be interested in seeing benchmarks for the MT variant – that really oughta give the PM a “run for it’s Watt”.
    Considering your results, it’s just funny that the Turion has gotten so litte press for it’s apparent performance as a mobile CPU. (Well, rather that “funny”, it’s probably a sign of Intel’s qualities regarding marketing campaigns. (And goes to show how much of an impact on public perception it can have if you can just spill out the $MM for advertising))

      • accord1999
      • 14 years ago

      How is the Turion close in power consumption? It uses in the neighborhood of 40+W, the P-M is ~20W. The Turion CPU uses more power at full load than most P-M laptops. Basically, P-M is to Turion ML, as A64 is to Prescott.

      And the idle power of a desktop system can’t be extrapolated to a laptop, considering a typical P-M laptop uses 10-15W in idle. Without knowing the actual power consumption of the individual components, you cannot say that the Turion has better idle power consumption. It’ll be hard to beat Dothan, which idles at <1W at 800MHz.

        • wierdo
        • 14 years ago

        The review explains it pretty well imho: most users have their PCs at low utilization most of the day so the actually power consumption should be about the same or sometimes even less at those casual usage levels. For me personally, though, that may not be the case, but I’m price sensitive so that advantage would not be enough for me personally at current prices should I go for a notbook setup.

          • accord1999
          • 14 years ago

          But these system setups are not representative of a laptop. A typical P-M laptop idles at 10-15W, with the CPU consuming <1W at 800MHz.

        • Koly
        • 14 years ago

        You are right, the tested Turion is not competitive power wise with P-M, the difference under load is substantial and other components are more important when idle. Not that it is a bad mobile CPU, it’s very powerful, but it eats more juice. It seems to be right at the very edge of stated 35W TDP. As this is the top dog, I am very curious how lower clocked Turions would fare, maybe they would be more competitive. I am certainly looking forward to the MT-42 vs. ML-42 comparison Damage announced, I hope P-M will not be left out.

    • swaaye
    • 14 years ago

    In my experience and testing, RMClock will slow you down slightly if it’s running in the background. Unfortunately I can’t think of a way to control C&Q’s voltage without using RMClock…

    Systool has a VID changer but I don’t remember if it’s permanent (until reboot).

    • flip-mode
    • 14 years ago

    Thank you for making me a secure AMD fanboy 🙂

    • Toasty
    • 14 years ago

    But why did TR wait 24 hours before releasing this information to the press? And why did Andy Brown choose to interview only with Fox News when the news broke that the Worldbench numbers were wrong? The delay regarding updated power consumption numbers and the like leads us to believe that this is one of the most secretive and disingenuous hardware sites in recent decades, and goes to show that they are willing to lie to the American public in matters both large and small. For shame, TR, for shame.

    If this were /. I would get a +5 (Funny)

      • FireGryphon
      • 14 years ago

      ROFL! Oh, man, that was great =)

      • Logan[TeamX]
      • 14 years ago

      BAHAHAHHAHAHAH!

      Pure gold. Love it.

    • PerfectCr
    • 14 years ago

    4 MORE YEARS! 4 MORE YEARS!

    • slymaster
    • 14 years ago

    I have started buying Turion notebooks from HP for users in the office. The power consumption numbers may be higher than the Pentium M under load, but under real world use, the battery life is amazing.

    I have only deployed four Turion models so far, but all four users have come back to comment on how good the battery life is. The Turion models compare very favourably to Pentium M models for battery life.

    The performance is very snappy. The notebooks do not have any problems with heat or noise – they are extremely quiet. When one considers price, I would not hesitate to recommend a Turion notebook.

    • Ricardo Dawkins
    • 14 years ago

    I think the Pentium M power consumption numbers besting Turion are based on the Centrino platform…so AMD fanboys remember this test are run on motherboard not intended for notebooks.

    On a laptop (Centrino case) even the wireless adapter can reduce power consumption. So what is the matter if the Turion is competitive power wise..when you bring it to mobile it eat more power than a Centrino equivalent ?…:P

      • Beomagi
      • 14 years ago

      Anyone notice that the pentium-m is doing extremely well, leading in all game tests (showing it to be a real nice gamer cpu) yet in the cpu test for a gamer benchmark, it lags behind?

      What’s the point of a gaming cpu benchmarks that reflects the opposite of every other game or standard (read: uses a vidcard) gaming benchmark?

      • wesley96
      • 14 years ago

      You do realize the argument can work BOTH ways? And come on, don’t automatically label some people ‘AMD fanboy’.. it really makes your argument weak.

        • tfp
        • 14 years ago

        If the shoe fits, I mean try reading some of the posts in here.

    • Hattig
    • 14 years ago

    Thanks for the update and for listening to us whingers 🙂

    It does show that Turion was a good match for the Dothan in terms of processing power, power consumption, and cheaper to boot. The fact that it needed more clock speed to achieve that isn’t an issue, although the fastest Dothan (2.33GHz?) would probably outperform the fastest Turion (2.4GHz?) because AMD aren’t selling a 2.6GHz Turion (yet).

    • Koly
    • 14 years ago

    That’s more like Techreport, thanks. The 5W difference is smaller than I expected, 19W difference under load is hard to ignore, even if some of it is caused by DDR vs. DDR2. I am very curious whether the MT series shaves off only another 5W, that would be quite dissapointing. Now it looks like the situation in the mobile sector might be opposite to the desktop one: while Intel understates the TDP of its desktop CPUs and AMD overstate it, this mobile AMD CPU is somewhere at the edge of stated TDP while Intel’s have a respectful reserve.

    • wierdo
    • 14 years ago

    Interesting results. I guess this confirms a review I read earlier about Turion’s power consumption… it’s actually pretty decent relative to competition.

    • FireGryphon
    • 14 years ago

    Thanks for the update. All is well in Reviewland now.

    • Dposcorp
    • 14 years ago

    Thank you for making the corrections, as it is nice to see a site take constructive criticisms, and use them to not get offended, but to actually admit their were some errors and correct them.

    Bravo to you, sir, and the entire TR gang.
    Nice to see Dr. Evil come out and be evil as well.

    Also, I am very happy to the AMD chip do even better.

    It offers all the P-M does, and then some with it being 64Bit.

    And yes, it will matter as a lot of us are now running 64Bit code and Vista is around the corner.

    Also, props for the Pulp Fiction quote.

    All posts in this thread MUST have a PF quote related to TR.

    TR: It is a Tasty Burger!

    • Shintai
    • 14 years ago

    This is what I love on TR. The ability to redo and fix.

    • DrDillyBar
    • 14 years ago

    Word

    • dragmor
    • 14 years ago

    Much Kudos for the update

    • just brew it!
    • 14 years ago

    Thanks for the update.

    I love this site! 😀

Pin It on Pinterest

Share This