P4 SETI performance question...

[Jeff7181]

Duvie sparked this in the CPU forums... I thought I remembered seeing someone here say that while crunching a single WU they averaged about 2:30 per WU, and with HT enabled and crunching 2 WU's, they averaged about 2:45 per WU, so the output was nearly doubled. I can't find the thread I thought I saw that in... can anyone with a HT enabled P4 verify the ability to crunch nearly twice as many WU's with HT enabled and running 2 processes?

 

[SilverBack]

What you'll see is a output that's about 55% more than running one instance.
You'll see improvement, but not double.

 

[Engineer]

It won't be double, but it will be more! 🙂

My 3.2GHz HT P4 averages 2:49 per WU running 2 instances and averages almost 17 WU's per day. I don't think that it would average 1:24.5 when running a single WU with HT off! 😉

Fire up HT and run 2 instances! 🙂

 

[Jeff7181]

Originally posted by: Engineer
It won't be double, but it will be more! 🙂

My 3.2GHz HT P4 averages 2:49 per WU running 2 instances and averages almost 17 WU's per day. I don't think that it would average 1:24.5 when running a single WU with HT off! 😉

Fire up HT and run 2 instances! 🙂

Try turning HT off and just running one and see how much time it takes, I'm interested to see the difference.

 

[Ionizer86]

Sounds like fun comparing the results. For reference, my XP chip is at 2600+ (166 * 12.5), and WU time is between 2:30 and 2:35 on average. That's (IIRC) about what my bud's P4 2.4 @ 3.0 gets with one instance.