Test setup

For our little DFI adventure we put together the following test system:

Test Setup
CPU	Intel E6700 Core 2 Duo Intel QX6700 Core 2 Quad Cooled by TTIC "Big"
Mainboard	Asus P5B-Deluxe/wifi AP (modded) DFI Lanparty UT 680i LT SLI-T2R
Memory	2x 1Gb OCZ PC8500 Reaper HPC 2x 1Gb OCZ PC9200 FlexXLC
Other	Silverstone TJ09 Temjin Tower Case BFG 8800GTX (SLi) LG DVD-rom drive Wester Digital Raptor WD740 SATA OCZ ProXstream 1kW Watt PSU

As you can see, we have used two different setups for our benchmark testing. The first is based around our guinea pig of the day of course, while the second is based around the "old and trusty" Asus P5B-deluxe. The rest of the components were kept the same in order to get comparable benchmark results.

Overclocking the Core 2 Duo E6700

Let's get started with the E6700. Intel's E6700 has a maximum multiplier of 10, which can be dropped to 6 when using speedstep technology, or when the appropriate options in the bios are available (as they are here). For my little OC test, I was not interested in getting the maximum speed out of this CPU, as this is largely bound to the cooling that you use. Using single stage phase-change cooling, this particular CPU has already reached 4.6Ghz on other motherboards, using water-cooling, it's possible to get it stable at 4Ghz, using high end air cooling just a little less. A repeat of these tests I didn't find very interesting, as the results vary on various occasions (not the least being the ambient temperature).

Far more interesting in my opinion is a front side bus test. On motherboards with an Intel chipset, the "full cache" Conroe's - those with the full 4Mb of Level 2 cache - had some difficulties with very high front side bus performance. Whereas 400-450 FSB was not much of a problem, 500FSB was very hard to reach. I for one have never been able to get the E6700 stable past 460FSB, no matter how high the voltages of the Northbridge was increased. The reason for this is explained in another one of Tony's excellent articles, where with the help of FCG he discovered the effect the lowering the multiplier of the cpu had on the chipset and its overclockability.

Luckily for all of us, the 680i chipset is an entirely different story. After lowering the cpu multiplier to 6x, and by tweaking the northbridge voltage and the cpu VTT accordingly, we managed to get the following - rather stable - maximum front side bus :


509Mhz FBS with Core 2 Duo, previous record with this CPU was 460Mhz

This bus speed is what I call "rather" stable, as the Orthos stress test would not continue longer than about 2 hours of testing. All other benchmark tests could be run without any problems with the above settings.

Benchmark suite

With both setups, and with both processors, we're going to run the following benchmark suite and put the results in a few nice graphs for easy comparison:

Sisoft S.A.N.D.R.A. 2007

Lavalys Everest 4.0

Maxon Cinebench 9.5

MadOnion/Futuremark 3D Mark 2001SE

Futuremark 3D Mark 2006

Superpi 1.5, Xtremesystems edition

Game : 3D Realms Prey demo ; benchmark by Hardware OC

Results of the DFI 680i vs Asus P5B-Deluxe next -->