Paul McHale wrote: > I have always heard ECC memory is more reliable, but there is a performance > penalty. Crucial.com estimates the penalty to be 10% to 15%. That's a lot.
Negative. My own memory test benchmarks using a (cough) windows based test program shows only a 1%-2% speed degradation. Of course that depends upon the chipset used. But I wouldn't expect anything above 5% in the worst case for anything reasonable. What I frequently read on the subject comes from the overclockers. They are usually wringing the last drop out of the machine and reliability running is not really a consideration since they are usually running MS. But the result is really a lot like this reference link. (Be sure to scan down the photos to the end or you will be missing out.) http://forums.vwvortex.com/zerothread?id=776885 Yes it goes fast. But is it your regular car? I don't think anyone doubts that those guys were serious about getting the performance they wanted. > There are some statements to the contrary with newer chipsets. I have the > 875P chipset and have no idea what the performance difference is. Any > recommendations or empirical performance data ? I have used a variety of AMD and Intel chipsets without seeing any huge penalty. Your best answer it to run some benchmarks and to determine this for yourself. A good memory test is memtest86. You can apt-get install it and it can even become a boot option which is very nice. But I have not used it to benchmark memory speed. Can someone suggest a good easy to use memory benchmark program? Bob
pgp00000.pgp
Description: PGP signature