Cannyone wrote:I have a slightly different "take" on this topic. Intel has been pushing it's partners to be more "energy efficient", which obstensively is a good thing. But last year I bought an Asus P5E (plain jane X38 chipset board) when another Asus board (a P5B-Plus) encountered a problem with the BIOS. Everything seem to work fine at first. Then I started getting BSODs and other strange errors. I finally traced the issue to memory problems and ran some tests. First off, the memory wasn't running at the speed specified in the BIOS. And since I was getting massive amounts of errors, I replaced the memory. When I got the new memory I ran the same tests and discovered the same issue with the memory's speed. My best guess was that this is a "feature" of the chipset, where until the system gets Windows running it does not run some of the components at "full speed".
It wasn't until a few months later that I discovered the full extent of these new "efficiency" measures. I took the P5E and the new Corsair XMS2 memory to a friend who does "failure analysis" for a living. He determined that the P5E was under-volting the memory. Specifically the default value is set at 1.8v, and the board was only supplying 1.74v. The original pair of modules were spec'd to run at 1.9v. But this board was consistently running all memory in that 1.74v - 1.82v, even when higher values where set up in BIOS. Now the real question is: is this something Asus is responsible for? Your article suggests to me that they are primarily responsible.
Now there isn't anything I can do about this situation. Except avoid purchasing Asus products, and advising people who really insist on purchasing the same to be careful of the memory they select to use in those boards. Specifically they need to be careful in what they expect to work. For example; I now have some OCZ memory that is running at 1000MHz, with 2.1v, it will run with 1.8v but only at 800MHz (I'm not using that P5E for my main gaming rig now. I switched to an MSI P7N Diamond {Nvidia 780i chipset} which runs this memory perfectly!). When selecting memory they should pay special attention to only using memory that Asus certifies to run with their board. Or risk premature failure of your memory, if you are running the memory at other than default values. Gigabyte doesn't seem to have these issues and I recommend their products highly.
Cannyone,
I'm not calling you a liar or anything like that, but that is not the case with this motherboard. I usually don't show voltage checks in the article, but I do actually use a digital volt meter to check the actual voltages on the board on all the motherboards I review. The ASUS P5E3 Premium does not undervolt the memory at all, it actually over volts it a bit. I re-built the test system used for this article and ran some VDIMM voltage tests just for you and got my girlfriend to take a picture of me doing one of the checks so you can see that I'm not bull shitting you.
- asus_voltage.jpg (80.86 KiB) Viewed 6056 times
Here is what 1.90 VDIMM set in the BIOS is really feeding the modules. I used my DMM to get the reading on the actual power feed pin in the DIMM slot itself, which is the last component on the board before the power is put to the memory module. I like this reading over all others.
AUTO: 1.632V (The BIOS starts at 1.50V and goes up, so auto is set to 1.6V)
1.8V = 1.830V
1.9V = 1.928V
In general ASUS seems to be over on the voltages and not under like you suggest. On average the memory is getting ~0.030V more than what you set in the BIOS no matter the setting. This of course is a different board than what you had, but I wouldn't say to avoid purchasing ASUS products. It is likely the board you personally had could have been defective, which then in turn caused the failure you saw. Either way you look at it, you got the bad end of the deal though ;(