Super PI and FAH

[marty9876]

Huh, just for kicks I ran that Super PI thing. With folding running, 57.x seconds @ 1M, without folding running 47.x @ 1M.

Just interesting, looks like folding did slow down something quite a bit. Not even running a QMD core, a 600 beta pointer.

huh

 

[Carnival Forces]

what's the process priority for SuperPI ?

if both are at "low" then i can see where the competition could be taking place...

console version of FAH, i'm assuming?

 

[marty9876]

process priority= both @ default. Super PI is "normal" with FAH at low or what not.

Is a HT box w/ one instance of FAH running, not sure if that matters

 

[unhappy_mage]

How much memory used by FAH? This is usually the culprit in my tests...

 

[marty9876]

The box has 2 gig, FAH ~100 or what not.

 

[Carnival Forces]

The box has 2 gig, FAH ~100 or what not.

that's really odd; i don't know what to make of it

you sure it wasn't another process running at the same time? try doing the two tests again in Safe Mode

 

[marty9876]

The box is clean, I'm running the long one now, was 32:x minutes w/o FAH, already over that and only 2/3 way through.

 

[Flying Fox]

My noobish explanation is that HT is indeed to blame. Since XP sees separate processes and 2 logical CPUs, it thinks it is safe to put the 2 active threads on the 2 "cores". But since HT is really not 2 physical cores, bottlenecks are bound to happen.

If you can, try disabling HT or set both processes to the same processor affinity, that should explain it.