Sign in with
Sign up | Sign in

Power, Heat, And Efficiency

System Builder Marathon, March 2011: $2000 Bonus Build
By

More powerful graphics cards hurt the new build’s power consumption numbers, but only when 3D applications are running. Fortunately, we also know they increase performance. So, the efficiency charts should reflect those two figures balancing each other out. These charts only reflect one half of that equation.

CPU temperatures are excellent for both systems. Because GPU overclocking required manual fan adjustments, the numbers are not comparable in a precise manner.

The new build outperforms the original in games, but nowhere else. The former build actually beats it slightly in overall performance.

We reduced the effect of hard drive performance on overall performance scores from our previous 25% to the current 10%, since most people spend no more than 10% of their work time waiting for programs to load. Our three non-gaming application suites account for 30% each of today’s total performance score.

Small graphics performance gains that usually occur only at our highest test settings were more than offset by increased power consumption in our efficiency charts, allowing the original build to take a big lead.

Display all 59 comments.
  • 4 Hide
    sandmanwn , March 30, 2011 4:09 AM
    More opportunities to win a free giveaway!
  • 3 Hide
    sabot00 , March 30, 2011 4:10 AM
    Yay! 1 more chance to win!
  • 1 Hide
    hayest , March 30, 2011 4:13 AM
    Great build! Can't wait to see what is in store for next quarter!
  • -1 Hide
    Anonymous , March 30, 2011 4:16 AM
    Great SBM series. What about an AMD quad core vs. an AMD triple core (unlocked 4th core) vs. Core i3 value comparison?
  • 1 Hide
    compton , March 30, 2011 4:28 AM
    Now this is pretty interesting. I wouldn't have suspected that the memory bandwidth reduction in the alternate build would have been as big of deal as it is. I think both builds are fantastic, but the NF200-equiped Asus board was one I myself was trying to obtain. It has been deactivated on the Egg, and the winner of the original $2000 build will be extra lucky indeed. Hopefully, it's me.
  • 1 Hide
    Crashman , March 30, 2011 4:37 AM
    comptonNow this is pretty interesting. I wouldn't have suspected that the memory bandwidth reduction in the alternate build would have been as big of deal as it is. I think both builds are fantastic, but the NF200-equiped Asus board was one I myself was trying to obtain. It has been deactivated on the Egg, and the winner of the original $2000 build will be extra lucky indeed. Hopefully, it's me.
    It was deactivated, but has since been reactivated:
    http://www.newegg.com/Product/Product.aspx?Item=N82E16813131714
    Now when you win the second PC, there will be no need for sadness: You can upgrade it with the motherboard from the first!
  • 0 Hide
    wolfram23 , March 30, 2011 5:01 AM
    I'd love to see more info on 16x/16x vs 8x/8x. I just looked at the last time you guys reviewed it, but it was only with a GTX480. Being one of many with a P55 8x/8x mobo, I really need to know what my best upgrade path would be... single top end card at 16x? Or CF/SLI again? Is the resolution going to be the deciding factor? Seems like high res sees less of a PCIe bottleneck at 8x, but maybe it's just due to framerates being lower? Need info! lol.
  • 1 Hide
    Crashman , March 30, 2011 5:09 AM
    Wolfram23I'd love to see more info on 16x/16x vs 8x/8x. I just looked at the last time you guys reviewed it, but it was only with a GTX480.
    Look for something in a week or two. And you're right to NOT bother with the GTX480 stuff, Nvidia cards need less bandwidth than AMD cards.
  • 3 Hide
    compton , March 30, 2011 5:09 AM
    Quote:
    It was deactivated, but has since been reactivated:
    http://www.newegg.com/Product/Prod [...] 6813131714
    Now when you win the second PC, there will be no need for sadness: You can upgrade it with the motherboard from the first!


    This is why the SBM is the best thing going. A few other sites do similar articles, but Tom's is far and away the champion. Another well respected site doesn't even build the systems, but Tom's builds three (or four!) and gives 'em away like sweet delicious candy. Every build has its's own quirks, issues, and performance wins (losses too) that can't always be understood until the gear arrives and goes together. If system building was entirely predictable, no one would build their own. It's just more fun this way.
  • 1 Hide
    qwertymac93 , March 30, 2011 5:32 AM
    antec 300 again... im really not liking that case. you have any experience with the haf 912? it seems like a very nice case.

    Anyway, i never understood why intel went with just 16 lanes on SB yet all the mobo makers market their ultra high end cross fire boards. :pt1cable: 
    Thing I'm wondering is, if intel switched to 24 lanes, could the graphics cards work at 12x each for 2 way, and 8x for 3 way? i know a full 32 lanes is unlikely, that's why I'm asking.
  • 0 Hide
    sparky2010 , March 30, 2011 6:02 AM
    i think i'll wait until the the Z67 is out.. will be interesting to see if quicksync will add anything performance wise.... and while i'm at it, will probably wait to see what AMD have in store with bulldozer.. i mean, i've waited 2 years, i think i can handle a few more months! but it's starting to get AHHHHHHHHHHHHHHH!!!
    yeah..
  • 0 Hide
    sudeshc , March 30, 2011 6:58 AM
    I liked this build too, its quite upto the mark.
  • 0 Hide
    rainwilds , March 30, 2011 9:01 AM
    Wolfram23I'd love to see more info on 16x/16x vs 8x/8x.


    Over at HardOCP they did a good review on just that. Both bandwidths perform the same for current cards. However they only tested up to 4xAA at 2560x1440. Might be different at higher AA.
  • 1 Hide
    silverblue , March 30, 2011 9:10 AM
    I can't help but wonder how things would've changed had those 6950s been unlocked. It might have negated the use for a follow-up article, I suppose.
  • 3 Hide
    iam2thecrowe , March 30, 2011 9:39 AM
    Question for the guys at Toms, have you ever noticed "microstuttering" in your recent crossfire builds and reviews? Maybe you should do an article on this and how to overcome it. Maybe a "Crossfire guide" or something showing settings to change and how to get things to work properly. microstuttering, incompatability, instability and poor performance with some games are always reasons people do not choose crossfire. Can the experienced people at Toms shed some light on this subject?
  • 1 Hide
    mattmock , March 30, 2011 10:28 AM
    "we doubt anyone building with these cards would choose anything less than 2560x1600"
    I am running two GTX 570s in SLI at 1080p on a 120hz monitor. Its overkill for most older games, but some newer ones push it a bit. Using this setup gives nice high minimum framerates at max settings+4xAA in games like STALKER:COP complete, metro 2033(sans ADOF) and just cause 2. Keeps things noticeably smoother than one 570.
  • 1 Hide
    dcompart , March 30, 2011 12:39 PM
    Thomas S., can you add a "1920x1080 resolution High End gaming Value performance/relative price" graph/illustration at the conclusion page in the future? I understand that these are enthusiast systems and part of the point of benchmarking the higher card 6970 system is to point out the performance increase at higher resolutions, but 2560x1600 monitors can go from $1000/2000+ on Newegg.com. Some people who splurge and build a $2000+ system may not have deep enough pockets to get anything higher than 1920x1080 and it would be great to have the informative percentage graphs for the suggested resolution.
  • 1 Hide
    krinkles619 , March 30, 2011 1:17 PM
    If i won either of the 2000 pc's i would put them in a NZXT Phantom, because hey, if you've got almost 2k worth of parts in it you might as well have a sexy case, amirite?
  • 0 Hide
    Onus , March 30, 2011 1:51 PM
    Six of one, a half dozen of the other...with a $2K budget, it is much more difficult to build a "bad" system, than it is with a $500 budget to build a "good" one. The tradeoffs, while having measurable effect, are small and pretty meaningless to the user experience. I am absolutely not implying this was a waste of time; this comparison I hope served as a very good example that will assist others to not waste time.
  • 0 Hide
    impulse89 , March 30, 2011 2:00 PM
    Awesome addition to the set. I would love to win any of these machines.
    I would have thought the 70's would have made a bigger difference in lower resolutions, but apparently not. Wish i had a 2560x1600 setup.
Display more comments
React To This Article