HardOCP hat das Teil im Review. Dort wird auch auf die Probleme hingewiesen. Ist noch ganz interessant. Unter anderem wird ein User aus dem HardOCP-Forum zitiert:
-----------------
Originally posted by BaWoofers
As I proposed on my first post in page 43 it is not the memory or the CPU that is preventing good overclocks. It is a weak Vcore power supply circuit on the motherboard. As people are now posting, they get better results at lower voltages, not higher ones. This is opposite of usual overclocking approaches. Another piece of evidence is the people that have fast processors that are failing at default frequencies.
Proof:
I first ran memtest at 133 MHz x 4 mem 1.625 V. It ran for 32 minutes without a single error. I then applied a 1/3 ohm resistor load to the Vcore line on the motherboard. This resulted in an additional 7 watt load on the power supply. Memtest then came up with 874 errors in the next 4 minutes that it took to complete the next pass. I then disconnected the load and memtest continues to run without errors, another 30 minutes so far.
Let me restate it. I went from working to failing to working without changing memory timing, CPU frequency or even rebooting.
People that are getting poor overclocks can test for the problem by finding the fastest stable frequency at 1.5 Vcore and then just increase Vcore by 0.1 volts. This has a similar effect as my test. Running at a higher voltage will result in a larger power draw and you will fail tests.
This also explains why you can get some programs to run at much higher overclocks. Different programs exercise the CPU differently and may draw less power. Does anyone know what program draws the most power? This can be inferred by which program raises the CPU temp the most. Windows XP puts the CPU in low power idle mode when processing demand is low. This may allow XP computers to overclock better for day to day use.
What's next?
I did take the board to work and scope various lines. The Vcore shows significant high and low frequency fluctuations. This matches the fluctuating readings everyone gets when monitoring it.
My first mod was to add a 10,000 uF capacitor to Vcore. This did not help. Next I replaced the 10,000 uF cap with seven 47 uF and two 100 uF tantalum caps. These are supposed to help with the high frequencies. This did not help. Lastly I added eight 47 uF chip tantalum caps on top of the decoupling caps right under the CPU heatsink. Still no improvement.
I still don't have a solution. I don't know what is so different that some people get much better results. My board, 1.6a and memory all came from Newegg.
I would not be surprised if there was a recall on the board.
No BIOS patch will fix this board.
-----------------
Fazit des Reviews:
-----------------
I had high hopes for this board diving into it. With the features and performance this board has shown us I was ready to give it a HardOCP Top Pick award. But with all the trouble I had with stability and overclocking this board has certainly not earned it. The board is laid out very well; they were able to put a lot into a little space. It has features galore, having everything you would need for an enthusiast motherboard. The support for 32bit RDRAM pushes it one step above all the rest. Its software package and manual are exceptional. But I have a very hard time recommending this motherboard with the problems present. If so many people are having problems, then something must be wrong, right? I don’t know what happened, but Asus introduced to use some very good technology, but this motherboard is definitely not made with the quality and stability from Asus they have been known for. At this point in time with current BIOS I can only recommend this board for people that like to experiment or are hardcore enthusiasts who like to test different technologies out. Who knows, maybe you will have good luck with yours, but it’s that question that may push this board away from a lot of potential buyers. If Asus CAN fix this problem with a simple BIOS update I very much implore them to do so quickly. Otherwise, this may be one motherboard you pass up looking at other 32bit RDRAM
-----------------