Sign in with
Sign up | Sign in

Tuning The Components By Overclocking

System Builder Marathon Bonus: Newegg Customer Choice PC
By

Unlike the motherboard that Don picked for his $1300 configuration (alright, we'll stop poking fun at him now), Newegg's customer-choice Intel DZ68BC features a full range of overclocking capabilities. The same goes for the Core i5-2500K processor that the vendor's customers favor. The caveat, of course, is that this machine costs a couple hundred dollars more than what Don had in his budget.

The DZ68BC Performance menu has only a handful of settings, instead guiding users to submenus for most of the changes we wanted to make. We did use its Host Clock Frequency control, however.

Customer choices aren't always the right choices, and we get our first example of an issue we'd find in one of our motherboard round-ups. Setting CPU core voltage to 1.35 V yields an actual 1.37 to 1.38 volts, which is too much variation to make it something I'd use, personally. A Low V-Droop setting keeps the voltage between 1.36 and 1.39 V under heavy and shifting CPU loads.

Each power limit is increased to its maximum setting, and each Intel Turbo Boost ratio dialed-in to 45x. Although 4.6 GHz seemed like it'd be stable after an hour of stress, the system eventually froze.

Intel’s DZ68BC was also resistant to memory overclocking. It locks up when we set DDR3-1866 data rates using a set of modules that we know works fine at that speed. Instead, we used the memory’s DDR3-1600 CAS 9-9-9-24 XMP value, rebooted, and manually lowered its timings to quicker 8-9-8-16 latencies.

EVGA's cards came factory-overclocked, pushing reference 822 MHz GPU and GDDR5-4000 frequencies to 850 MHz and GDDR5-4104. We pushed even higher to a 945 MHz core clock and GDDR5-4360 memory with little effort. Most of my time was spent stability testing using MSI's Kombustor utility, 3DMark, and DiRT 3.

Display all 75 comments.
Top Comments
  • 20 Hide
    Crashman , March 30, 2012 5:23 AM
    mikenygmailNice, but would have been better with 2 x 6970 2 GB or 2 x 6950 2 GB unlocked to 6970.
    Better choices outside of "consumer choice" are irrelevant to a "consumer choice" selection.
    a4mulaInteresting, still not surprising given recent results in group dynamic studies. Groups will often make better choices than individuals, that's not to say a group can replace or perform on par with an expert individual, just better than the average.
    Yes, the motherboard could have been better AND cheaper if not for the fact that it was picked by the group rather than an expert individual, but the complete unit was still acceptable.
    aznshinobiJust saying, Asrock Extreme3 Gen3 Z68 would save some and not bad rated. Could go for something like the OCZ ZT 750w which also is solidly rated, both saving money and offering similar if not better performance. For the GPUs, probably would've been better just to go with 2x7850 just to see how it does after all I haven't seen a whole bunch of those benchmarks.Understood that this is best on best rated components, just saying it would've been nice to see the ones I mentioned for a value build.
    Right, part choices were limited to the top two rated parts, based on which of the top two customer rated parts most closely matched the rest of the system.
    DarkersonNot a bad system at all. Im just waiting for people to start whining about 680s like in the other builds.
    LOL, I'm waiting for a stream of "Why didn't YOU pick THIS" when Newegg Customers were the pickers and the "THIS" they're screaming about doesn't even have a customer rating :) 
  • 19 Hide
    Pezcore27 , March 30, 2012 4:47 AM
    Is it sad I liked this build the best out of them all?
  • 12 Hide
    a4mula , March 30, 2012 4:49 AM
    Interesting, still not surprising given recent results in group dynamic studies. Groups will often make better choices than individuals, that's not to say a group can replace or perform on par with an expert individual, just better than the average.
Other Comments
  • 19 Hide
    Pezcore27 , March 30, 2012 4:47 AM
    Is it sad I liked this build the best out of them all?
  • 12 Hide
    a4mula , March 30, 2012 4:49 AM
    Interesting, still not surprising given recent results in group dynamic studies. Groups will often make better choices than individuals, that's not to say a group can replace or perform on par with an expert individual, just better than the average.
  • -1 Hide
    aznshinobi , March 30, 2012 4:57 AM
    Just saying, Asrock Extreme3 Gen3 Z68 would save some and not bad rated. Could go for something like the OCZ ZT 750w which also is solidly rated, both saving money and offering similar if not better performance. For the GPUs, probably would've been better just to go with 2x7850 just to see how it does after all I haven't seen a whole bunch of those benchmarks.

    Understood that this is best on best rated components, just saying it would've been nice to see the ones I mentioned for a value build.
  • 3 Hide
    stm1185 , March 30, 2012 4:58 AM
    mikenygmailNice, but would have been better with 2 x 6970 2 GB or 2 x 6950 2 GB unlocked to 6970.


    Yeah that was what I was thinking, if you have a 2560x1600 monitor then the 2 6950s wont see that performance hit at that res like the 560ti's do. And would outperform the 7970 as a result while still costing enough less to move up to that 2500k.

    When I can drop $1300 for a Dell U3011 or HP ZR30w I doubt I would be pairing it with a $1300 PC, so I wonder if its even necessary for a mid range build as how often are you really going to find that pairing. Though hopefully soon Apple is going to push the LCD makers kicking and screaming into the 4k and 8k display era!


  • -2 Hide
    hmp_goose , March 30, 2012 5:13 AM
    Wow: And here I was thinkin' "maybe the SBM should work like the monthly Best X articals, where it's not anchored to a price point".

    Drunk Min's t'ink alac, and all that.
  • 1 Hide
    ammaross , March 30, 2012 5:19 AM
    "...with none of the compromises that plagued Don't maligned build"

    Last page. Should be "Don's" but "Don't" works good enough :p 

    Definitely would have loved to see a pair of 2GB cards duke it out though. The base system called for it (nearly). Take the price from the hide of the mobo.
  • 6 Hide
    Darkerson , March 30, 2012 5:23 AM
    Not a bad system at all. Im just waiting for people to start whining about 680s like in the other builds.
  • 20 Hide
    Crashman , March 30, 2012 5:23 AM
    mikenygmailNice, but would have been better with 2 x 6970 2 GB or 2 x 6950 2 GB unlocked to 6970.
    Better choices outside of "consumer choice" are irrelevant to a "consumer choice" selection.
    a4mulaInteresting, still not surprising given recent results in group dynamic studies. Groups will often make better choices than individuals, that's not to say a group can replace or perform on par with an expert individual, just better than the average.
    Yes, the motherboard could have been better AND cheaper if not for the fact that it was picked by the group rather than an expert individual, but the complete unit was still acceptable.
    aznshinobiJust saying, Asrock Extreme3 Gen3 Z68 would save some and not bad rated. Could go for something like the OCZ ZT 750w which also is solidly rated, both saving money and offering similar if not better performance. For the GPUs, probably would've been better just to go with 2x7850 just to see how it does after all I haven't seen a whole bunch of those benchmarks.Understood that this is best on best rated components, just saying it would've been nice to see the ones I mentioned for a value build.
    Right, part choices were limited to the top two rated parts, based on which of the top two customer rated parts most closely matched the rest of the system.
    DarkersonNot a bad system at all. Im just waiting for people to start whining about 680s like in the other builds.
    LOL, I'm waiting for a stream of "Why didn't YOU pick THIS" when Newegg Customers were the pickers and the "THIS" they're screaming about doesn't even have a customer rating :) 
  • 4 Hide
    Darkerson , March 30, 2012 5:27 AM
    Im sure it will be inevitable. ;) 


    Stupid TomsHardware, Y U NO PICK MY PARTS! :p 

    Edit: Obvious sarcasm is obvious. Ah well, Ill take this as all the non article reading 680 noobs being offended. I dont care.
  • 4 Hide
    jezus53 , March 30, 2012 6:35 AM
    So in the table under software for the $1300 PC you put "Nvidia GeForce 285.62" for the graphics driver. I'm hoping that's a mistake because you have a 7970 installed! :p 
  • 4 Hide
    EzioAs , March 30, 2012 6:46 AM
    Nice article. I personally prefer Don's build over this one since he focused more on gaming(without relying to multi-gpu setups) and his build had better efficiency although this one has a better power supply, better case, cpu cooler and a higher capacity ssd.

    To me, multi-gpu setups are just a cheaper alternative to get better performance but introduce many issues that Tom's Hardware always states(micro-stuttering, heat, power and scaling), it's definitely better to get a more powerful single card. Not to mention the 560ti in this build only has a 1GB of framebuffer, which is only enough for max details at 1920x1080/1920x1200 but once your target is 2560x1600 or triple monitor gaming, it fails just as shown in the article.

    For most gamers out there, a 64GB ssd is probably enough for windows and programs, and let's face it, loading time doesn't dramatically decrease when you install your games on an ssd. Yes, they do loads faster but definitely not enough to justify paying more just for games. If you want it to load faster, caching using intel SRT on the Z68 platform is the better choice.

    Overall, I agree with the last statement on the article, "the customers are mostly right".
  • 8 Hide
    lightbulbsocket , March 30, 2012 6:55 AM
    I'd be really curious to know if and how much better a pair of GTX 560 Ti 2GBs would have done in this comparison.

    That's one thing I haven't really stumbled across yet is a good test for how a pair of GTX 560 Ti 1GB cards compare to the 2GB cards. Has anyone done a test like that?
  • -4 Hide
    executor2 , March 30, 2012 6:55 AM
    @EzioAs " Yes, they do loads faster but definitely not enough to justify paying more just for games" . Are you f*cking serious ?
    Load times half with using my SSD , how about 1 minute waiting in Dragon Age halved to 30 seconds ? , how about instant loading of application , how about INSTANT alt tabbing though the games .. come on man , don't speak if you don't have a SSD ( or have money for it ) . 64G SSD are CHEAP , and anyone who has a salary can afford them.
  • 6 Hide
    EzioAs , March 30, 2012 7:10 AM
    Quote:
    @EzioAs " Yes, they do loads faster but definitely not enough to justify paying more just for games" . Are you f*cking serious ?
    Load times half with using my SSD , how about 1 minute waiting in Dragon Age halved to 30 seconds ? , how about instant loading of application , how about INSTANT alt tabbing though the games .. come on man , don't speak if you don't have a SSD ( or have money for it ) . 64G SSD are CHEAP , and anyone who has a salary can afford them.


    No need to be rude, but seriously, it depends on the game. Some loads faster quite significantly, but other don't and for a gaming oriented system, SSDs are very optional. Maybe you don't fully understand what I posted before, but I did say 64GB is enough for gamers, not games. What I meant is the required capacity for the OS and other programs but with only 64GB, it's probably only enough to install 1-2 modern games depending on the size. Games can still be installed on the hard drive and on the Z68 platform, you can use Intel SRT to use the ssd as a cache for the hard drive. That way, your most commonly used programs or games still has the quick responsiveness as when you used an ssd.

    I stand by on what I said before and I apologize if I offend you in any way
  • 5 Hide
    Crashman , March 30, 2012 7:14 AM
    lightbulbsocketI'd be really curious to know if and how much better a pair of GTX 560 Ti 2GBs would have done in this comparison.That's one thing I haven't really stumbled across yet is a good test for how a pair of GTX 560 Ti 1GB cards compare to the 2GB cards. Has anyone done a test like that?
    It might have also been interesting to compared only #1 picks rather than #1 and #2 combined. A pair of GTX 550's, 64 GB SSD and only 4GB of RAM? Probably only the games would suffer.
  • 2 Hide
    Goldengoose , March 30, 2012 7:59 AM
    Great idea. nice looking build too, seems the community go for the flashiest rather than always the best performer.
  • 3 Hide
    esrever , March 30, 2012 9:51 AM
    this is a great PC! I'd rather have a 7950 but thats just me.
  • 9 Hide
    mistyirc , March 30, 2012 11:45 AM
    It's worth pointing out that customers will factor in things that don't show up in benchmarks - for instance, the reliability of Intel motherboards or EVGA's customer service. Not to say the other PC's parts are sold by crappy vendors, but pure performance isn't necessarily the whole story of this PC.
  • 6 Hide
    Onus , March 30, 2012 12:02 PM
    Ok, but when Don RMA's his mobo and gets one back that works, he gets a "do-over." After all, the value comparison piece did mention that would be done before the parts were sent to the winners, so obviously it will need to be tested. While this "do-over" may not merit a whole benchmark article, I would definitely appreciate a short one calling out the performance differences.
    It looks to me like overall, Don's PC wins. At lower resolutions and/or settings, the CC PC has higher FPS, but both have FPS sufficiently high as to not matter. Now look at where the CC PC fails, such as BF3 on Ultra. FPS in the teens is not playable, but Don's PC is merrily fragging along, over 40FPS even at stock. And, Don's PC uses a lot less power doing it.
    Where Don's PC fails though is that miserable Apevia case; I don't want "that" under my desk. My 12-yr old nephew would probably love it, but it just isn't for me.
Display more comments
React To This Article