SBM: Overclocking The Competitors

[muk]

http://www.tomshardware.com/2007/09/21/sbm_overclocking_the_competitors/index.html

Armed with our low, mid, and high end systems, we do a little overclocking and see how each scales when pushed to the limit...

 

[magnus91]

Nice. I've been waiting for this.

 

[No1sFanboy]

Well done, good read.

I'm glad to see the all of the overclock details (voltages, timings) laid out. This kind of thing is a great resource for someone looking for some baseline numbers to work from when planning and executing their own overclocks.

 

[fumuthabeotch]

Great article, I like how all the details were laid out as well. Also great how the video cards were overclocked, most sites would've left it at just the CPU and RAM.
Too bad you guys couldn't get your hands on a G0 stepping Q6600 to make the mid-range even better. O well, still got the B3 pretty damn high!

 

[autoboy]

Before you start bashing your 450W power supply, saying it can't handle the load, why don't you plug it into a power meter and measure the power it draws so you can actually make a real statement about the system, rather than false conjecture? I bet your low end overclocked system ran barely over 325W at load. Still, an interesting read.

Tomshardware has been known to make fasle statements based on conjecture and myth. Well, this is a really easy one to fix. Measure the damn power usage and actually prove your statement. If you do that with every assumption you make, Tomshardware may be able to gain its reputation back.

 

[cleeve]

It IS conjecture. I don't know why you'd even assume I'm pushing it as a fact.

Clearly the existance of the word 'perhaps' is the sentence "Perhaps the PSU was struggling to provide fuel to both CPU and graphics card when they were drawing at the same time..." should have made it obvious it's a hypothesis.

If you want to take my conjecture as fact, well, go to town.

Regardless, watts at load wouldn't tell us much. It's amps on the separate rails that would indicate if we were pushing too far...

 

[nurn]

Wow, take a look at the difference between Oblivion Benchmarks at 1680 by 1050 for the high end vs the medium cost system. Gotta wonder if it would have been better on the midrange system to skimp on processor and mobo (throw in a Pentium E2160 for example, and overclock the heck out of it), then try to find a way to either SLI or Crossfire a couple of decent video cards. Say for example a couple of 512 meg HD2900XT's in Crossfire.

 

[cleeve]

For a gaming system, you'd probably get almost the same gaming performance with the e6750 overclocked paired with two 8800 GTXs in SLI.

It's the SLI'd 8800's that make the gaming difference in the high-end box...

 

[royalcrown]

I bet that the low cost sytem destroys the others on value, it's like drag racing with four times the horsepower to go just a little faster. Personally, I'll take a 1000 to 1500 dollar system and buy more software or nicer stuff to use versus 4000 dollars and less to go with it.

 

[Luscious]

It's unfortunate indeed you couldn't your hands on a G0 Q6600, it would have shown better scores squeezed the value of the high-end system even more.

 

[stemnin]

Most people that i've seen with 4k$++++ rigs were using CAD/3D/Photoshop and all that jazz, which is very expensive software (I think Photoshop is cheap at only 1k$?), the computer is a fraction of what they're paying and going to make (if they use it as such).

I wish my company would invest more in PCs, but you know old construction guys, all you need is CAD (nevermind explaining anything outside e-mail to 'em).