Hyperthread on X5DAE

Daytripper offers some great advice. Benchmarks can be a pretty bad way
of evaluating system performance. For one thing, they rarely make
proper use of advanced features such as Multi-CPUs, Hyperthreading, 64
bit PCI buses, etc... Hence, they give a false indication of system
performance. In addition, it's fairly common that one little tiny
problem can throw off the entire benchmark. I like real world
benchmarks. Fire up whatever application you're using and time a few
things with HT on and with HT off. Whichever turns out faster should be
your choice. If whatever you're using runs so fast you can't time it
accurately, send me your Dual Xeon and I'll ship you something that's
more appropriate. A better, non smart aleck way of saying that is:
If your application runs in 4 seconds with HT and 3.8 or 4.2 seconds
without HT, quit worrying about it. However, if you're encoding a video
using Premiere, and it takes 6.5 hours with HT and 7.3 without, stick
with HT...

Ed

Lee Johnson wrote:

 > Gang,
 > I enabled hyperthreading on an X5DAE equipped with 2 3GHz Xeons. However,
 > I ran PassMark Software's Performance Test with HyperThreading on and off.
 > It showed MUCH better performance in the test with HyperThreading turned
 > OFF. Why is this, and should I run with it off?
 >
 > Thanks
 > Lee Johnson
 > holeejo.TakeThisOut@bellsouth.net<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Hyperthread on X5DAE