Overclocking and Stability
In order to test the CPU voltage droop using digital PWM, we first used an E6750 G0 dual-core processor and found using a default 1.35V it idled at an elevated 1.384V in Windows. When we loaded the processor, the CPU voltage stayed firm between 1.384V and 1.392V, even when left for a while to allow everything to heat up. This is without doubt an excellent result, but for quad-core however, things were a little different. We dropped in a QX9650 and overclocked it up to 450MHz FSB at 1.46V and when it arrived in Windows we found this translated to just 1.376V which stayed the same under load.
So while the voltage is stable, it looks like you have to be a very liberal with it when overclocking a quad-core processor. In this respect we upped the voltage in the BIOS to 1.5+V when going further.
The QX9650 would hit 495MHz FSB just fine, at 1.7V north bridge, 1.51V CPU and 1.45V CPU VTT but crashed on POST at 500MHz. In contrast our old, yet ever capable X6800 sailed through 500MHz FSB at similar voltages however it locked up when we applied load to the system. It wasn't until we dropped it down to 490MHz FSB that it was fully stable.
Stability at BIOS defaults was excellent and the board didn't buckle under the load we put it under for prolonged periods of time. The heatsinks got warm to the touch, but never too hot and the whole machine just kept ticking over.
Value and Conclusions
Where the Asus Maximus Formula had a distinct lack of any performance advantage over P35, it at least offered a kick ass feature set on the board and in the box. The Abit IX38 QuadGT does make for an awesome board and BIOS, but it has a virtually non-existent supporting package and performs
worse than its little brother - the IP35 Pro. The features leave me completely polarised:
- The BIOS is absolutely lovely to use but its use of DDR2 is not at all optimised
- The board looks fantastic and the layout is great, but I don't like the inclusion of an x4 slot over having at least one or two PCI-Express Gigabit Ethernet sockets
- The cooling works well and it's got a pretty simple heatsink arrangement, however they are not screwed down like others and the PWM heatsink overlaps the rear I/O, meaning there are only four USB ports available
- If you need an extra x4 slot on top of CrossFire, then great, but it's a sacrifice at the cost of other features
- The on-board power, reset and clear CMOS buttons get an upgrade but the ones on the Asus Maximus are still better and far less flimsy
There are good, niche features like digital power regulation and fancy new on-board switches but as a whole for what you get, it isn't great value. On paper, before I had looked at the board I was really quite excited by the Abit IX38 QuadGT as it looks like the more attractive deal than the Maximus Formula - it has similar features and it's around £15 cheaper, but now after testing it the Formula is
even better value in retrospect.
I like the board in the same way that I liked the Maximus Formula, but its problem is that a six month old P35 board that costs a third less performs just as well, while a basic X38 board with the same core features will save you £20. Again, I'm just finding it hard to justify the extra cost.
Final Thoughts
The Abit IX38 QuadGT is a very solid and stable motherboard that overclocks very well and, as a result, it might appear to be an attractive proposition. Sadly though, it's let down by poor performance, a poor supporting package (which may not be a concern for many) and some of the board's features really polarise my opinion of it.
In the end, I find it hard to justify the IX38 QuadGT's asking price, but if you aren't fazed by the below-par performance and occasionally strange choice on the feature list, the board is solid enough to stand up to a barrage of abuse.
- Performance
- x
- x
- x
- x
- x
- x
- -
- -
- -
- -
- 6/10
What do these scores mean?
Want to comment? Please log in.