• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Multiple Instance on HT Question

Oldie

Oldie

Mean Old Administrator
2FA
Joined
Jan 12, 2004
Messages
21,915
How do you balance this out? Obviously it depends a bit on how much ram you have, but say I've got a 3.0 Prescott at stock speed (hypothetically of course) and a gig of ram. Do you know if there is any performance gain in doing 2 Large WU, 2 Small WU, or 1 & 1?

Currently I've got my HT's doing one large and one small. What do you guys think?

 
You don't really get to choose, you can only configure your client to accept big WU's. It's the luck of the draw. The Stanford servers deal out the WU's, according to a schedule that scientists decide upon.

With HT machines, it is better (point wise) to run 2 instances and some say it is a 15% improvement. But keep in mind, that Stanford has stated that they prefer one folding client to run on one cpu; because they like to get the WU's back fast.

Ciao for now .....
GwilyaGwees
 
Actually you could almost choose if you select the Timeless Tinkers for one process and -advmethods for the other.
However, provided a stable 24/7 system, processing the largest WU (QMD if you can get them assigned) seems to give the best bang for the cycle presently.
I'm getting about 1,000 points per week per Ghz on a system running the QMD's.
 
You could run two of the bigpackets Gromacs WUs with that setup and not have to worry about running out of RAM depending on what other software you are running. If you use anything else that is RAM intensive on a regular basis I would be careful about running to instances of those. The main thing would be to look to see how much RAM you normally use. I would fire up whatever software you normally run (like games) and after using it for a bit, check the peak memory usage using Task Manager. If the peak commit charge is well under the amount of physical RAM you have (this includes running both instances of the client as you have it setup with one bigpackets WU being crunched) then you could probably run two instances of bigpackets. Much of it depends on your system and RAM usage. with anything less than a gig of RAM I would not suggest running two instances of bigpackets WUs. A gig or more of RAM will depend on your usage.

Hope that helped.

 
I can easily get away with 2 large WU's on most of my HT boxen, I was just wondering if anyone saw a performance difference with any of the combinations. I guess I'll test it out for awhile and report back. ;)

 
If you have the resources to run two instances with bigpackets and don't mind the extra wait times to complete them, I would see no reason to not do it except for the reason Vijay put out about wanting to get WUs returned as soon as possible. He didn't say you shouldn't do it, just that he would prefer if you did not. As long as both make deadline in plenty of time, I would not feel bad about doing it.

 
You must log in or register to reply here.
Back
Top