Recommended reading

System Builder Marathon, May '09: $1,300 Enthusiast PC

Features
By published

Benchmark Results: Productivity

In the past, 3ds Max has sometimes shown us inexplicably large performance increases when systems were overclocked. This isn’t the case today, with a 26% speed increase that falls right in line with the media encoding applications we’ve tested so far.

AVG demonstrates a surprisingly low 3% performance increase when the system is overclocked.

WinRAR delivers a solid 31% performance increase with the higher clock speed.

WinZip is showing us a 25% increase in performance with the over-clock, which is par for the course.

At this point, we can safely assume the 2.66 GHz to 3.44 GHz Core i7 920 overclock will provide about a 25% performance increase in the average application, with a few exceptions.

Now, we can get to the heart of things: game benchmarks. This is what a portable LAN-party-sized system is made for in the first place, and is unfortunately where the CPU overclock will likely deliver diminishing gains as the resolution is raised and the graphics cards become the bottleneck.

110 Comments Comment from the forums
  • jtnstnt
    It looks like you guys have a fetish for silverstone sff cases, and small motherboards. You guys probably like other small things...
    Reply
  • one-shot
    Was the peak power draw measured from the wall? If so, do you know the efficiency of the PSU to determine the approximate power draw from the components at a given level of output? Overall, great article. I hope my 650Watt PSU with 3 X 19A 12V rails can handle another GTX 260 Core 216.
    Reply
  • serifus
    if you wanted to go SFF it would be nice to see one done in an LIAN-LI PC-A05NB. at least you still get the full atx boards in those.
    Reply
  • SpadeM
    This system is no slouch and cuts through our new gaming benchmarks like a hot knife through butter.
    I wonder how exactly does the selection of components go. I mean it seems that there's some attention given to the forums to be politically correct, but that's kind of it. For $1300 a Phenom 2 + micro AM3 board + 2x4890 in crossfire is a much better solution so .. why not choose the better option?
    Reply
  • capttylor34
    That system looks quite crowded on the inside, and that Dark Knight is big for a regular sized system, I'm actually kind of impressed you managed to fit it in a Micro ATX at all. I suppose as long as it runs stable, theres really no problem. Still think the articles should be titled "Micro System Build-off" just to let people know off the bat what they're about to read.
    Reply
  • IzzyCraft
    No space at all between those 2 cards. poor cards have no choice but to overheat.
    Reply
  • IzzyCraft
    SpadeMI wonder how exactly does the selection of components go. I mean it seems that there's some attention given to the forums to be politically correct, but that's kind of it. For $1300 a Phenom 2 + micro AM3 board + 2x4890 in crossfire is a much better solution so .. why not choose the better option?
    Reply
  • armistitiu
    I'm tired of seeing I7 920 in every damn "recommended" PC. I'm not saying go AMD but please just try to vary them a bit it's getting stupid. Also try building your own benchmarks or at least change them once in a while.
    Btw i\m getting tired of people picking up I7 and saying "because i do heavy CPU tasks" (not necessarily IzzyCraft ) and in fact all they need is a browser,OpenOffice and WoW minimized in the taskbar.
    The article is well done no doubts but try using other brands also. It\s starting to sound biased.

    Reply
  • ohim
    This article would be great if you could do : what you can buy for 1300 Intel parts and what you can by for 1300 AMD parts ... and put the 2 PCs to compeat to eachother :P it would be a more interesting article than just buy a pc and ... show what ?
    Reply
  • nerrawg
    Nice slick system this - looks like 2 x 260's is pretty good value at 1920 x 1200 for those that want to max out the settings in the most demanding games. Was wondering however if there is a good technical explanation for the power consumption results because they seem pretty strange?
    Reply