|by Eldonko | October 21, 2008|
Ok, now for the page you have been waiting for, let’s see how the DFI Lanparty JR P45-T2RS overclocks!
The board works great at stock, uses one of the latest and greatest chipset, has one of the most advanced BIOS on the market, and sample board results posted around different enthusiast forums show some nice results. However, all that means nothing if the board does not overclock well.
To cut right to the chase, we had a lot of trouble trying to overclock this board and we can say it certainly is not user friendly overclocking for a novice user.
First off, after seeing users post results running the JR at 600FSB and higher we were quite excited to test a high FSB board. We started by trying to boot at 500FSB, should be easy right? Ha, not even close. By using the standard BIOS settings we were unable to even get the board to POST at 500FSB; basically, it thinks for a second and shuts off totally and requires a CMOS reset and reset of the PSU to come back to life. Talk about hassle and if it weren’t for CMOS Reloaded, tweaking would still be in progress!
After moving into the advanced settings by pressing F9 on the main BIOS screen (and after many more hours) we were able to boot at 500FSB by tweaking TRD, GTL settings, and CPU and memory skews. The issue is VTT, GTL REF, PLL, skew are all metrics controlled by the CPU itself, and must be in perfect balance or you do not get any degree of stability, simple as that.
So we found settings that appeared stable (FINALLY) and OCCT was running at 500x8 (4Ghz). No big deal right? 4Ghz and 500 FSB is easy on many boards with a 45nm duo....well not quite. The test was set for 5 hours so returning to the PC later we were greeted with a corrupted Windows install. This was not repairable so Vista had to be reinstalled.
Many more hours passed and we were unable to get 5 hours of OCCT stability at 500FSB, the closest we got was 4 hours. As a result, we must conclude that the max fully stable FSB is below 500. However, benching above 500 is possible so we have included a shot of 32M at 520FSB.
Next we moved on from max FSB to shoot for a max overclock. Here there were some somewhat surprising findings. First, less voltage is better on our test board. It is hard to say if we were just unable to find the proper combination of settings at higher voltage for long-term stability or it is just not possible on this board. What could be a cause of the instability is the fact that this board uses 4 layer pcb and 4 a phase digital PWM, hardly components of a top clocker at high voltages.
Testing memory independently, the board was actually pretty good. Using Crucial Ballistix PC8500 (the new single sided stuff), 500Mhz (DDR 1000) was achievable at 4-4-4 and 600Mhz (DDR1200) at 5-5-5 at 2.15 to 2.25v without much trouble at all. All of the secondary settings you can imagine are available in the BIOS and the auto settings are set up quite well for users that do not want to tweak at that level. It is recommended to be cautious in interpreting memory errors properly though as an improper mix of GTL and skew settings can lead to memory errors as well.
In spite of the excessive tweaking, reboots, corrupt Windows installs, and headaches, we were able to achieve an overclock of 4066Mhz; a gain of 36%. To accomplish this, low vcore / VTT and the 400/800 strap were used, giving RAM speed of 453Mhz (DDR906) and timings of 4-4-4-4. Benchmarks were possible at 4100Mhz or so, but for the five hour OCCT test to finish we had to reduce to 4066Mhz.
It is important to note that a lot of time went into achieving stability, more than most would be willing to spend. The issue is the optimal setting combination may be completely different depending on ones CPU and what FSB you want to run at. Also what works for one person most likely will not for someone else; the only way to find what will work is some background on what each setting does and trial and error testing. These clock speeds will be used throughout the benchmarking sections.
Is this disappointing? Yes, to a certain extent it is considering the high (and obviously misplaced) expectations we had for the JR P45-T2RS based on previous experiences with DFI boards and numerous forum results we have seen. This goes to show two things: that not every board is made equal and a major purchase like a motherboard should never, ever be based on a few examples of extreme overclocks.
|Latest Reviews in Motherboards|