• 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.

?? Intel HT how to use 100% cpu?

fenderltd

[H]ard|DCer of the Month - July 2007
Joined
Feb 10, 2004
Messages
2,432
I have a few boxes that are HT that only tie up 50% cpu use. I would like to turn it in to 100%. Will I have to do a multiple client install for this to be possible? Or is it best to just run 1 instance of F@h Per HT Machine.
Thanks in advance!


 
Last I heard running 2 instances (the only way to get 100% CPU usage on HT P4's) only gives around 10-15% points boost versus running only one instance, so if you do heavy multitasking on it, like to play games or encode movies or something of the like, I'd recommend only running one instance. Otherwise, in order to save RAM (esp. if you're getting hit with QMD's), run timeless units on the HT CPU.

Edit: forgot badge.

 
fenderltd said:
I have a few boxes that are HT that only tie up 50% cpu use. I would like to turn it in to 100%. Will I have to do a multiple client install for this to be possible? Or is it best to just run 1 instance of F@h Per HT Machine.
Thanks in advance!



Actually with one client you are using 100% it just doesn’t show it.

You can in fact run two clients but you will only gain a 10 to 15% increase in output.

If you decide to run dual clients just follow the directions at hardfolding.com and don’t forget during initial set up to make two folding folders and in the config setup make one client machine #1 and the second machine #2.

I run a lot of duallys and I use the –local flag along with –forseasm and –advmethods.

Fold on.
 
Ok thanks for the info. It seems about 5 months ago I folded a good bit for BFG Tech and it seems like WU's went a hell of a lot quicker. Now it seems even on my fastest boxes F@H is dragging ass. Very slow. I am using the latest download I think 5.04? Or maybe I am just crazy!
 
Though it has been said time and again from Vijay and the crew at stanford that they would heavily prefer the HT people to only run one instance, as for SCIENCE the faster turnaround is more beneficial to them.


 
fenderltd said:
Ok thanks for the info. It seems about 5 months ago I folded a good bit for BFG Tech and it seems like WU's went a hell of a lot quicker. Now it seems even on my fastest boxes F@H is dragging ass. Very slow. I am using the latest download I think 5.04? Or maybe I am just crazy!
make sure you use the -forceasm flag. There are some MONSTER sized WUs out there, too. I'll also add my voice to the chorus telling you to use only one instance.
 
recently, the WUs have gotten bigger and bigger... Stanford had released new cores, new protiens,etc to take advantage of faster PCs...

As said above, make sure you're running the right flags, and that nothing else is stealing the CPU away (spyware, etc)....


Keep on Folding!! For the [H]orde!!

 
Through some not very exhaustive testing, I have found that to get anywhere near 10-15% extra in points you would have to run a gromac on one instance and a tinker on the other. If you run two gromacs core units, you won't see much, if any gain. If you run a QMD and anything else, it will actually net you less points. I've done a bit of experimenting with my rig here at work and it's a 2.8Ghz Preshott. Let's just say I only run one instance with QMD's on it.

Before I started the QMD's I ran one instance with bigpackets and one instance with timeless tinkers. The console running timeless tinkers was only set to use 15% CPU usage which would amount to the approximate gain in points I could see. No, the math probably doesn't work out right but it's close. This way, the normal units would be done in about the same time as if there was only one instance running plus I got the advantage of a few extra points. It worked out well all around. Also, increasing the percentange of CPU time on the timeless tinker console did not gain me any more points and slowed down frame times for the console running the normal units.

On the 2.8, the second console running Tinkers at 15% would process the 241-249 point Tinkers at about one frame every 64-67 minutes. That's the speed a PII 266 manages on those proteins.

There's a little data I have managed to remember from my testing with it.

 
If I am running the graphical client where can I add flags? Or Can I?
 
Actually... the general consensus around here is to run the console version.... then you can set everything up as a service, you'll never know it's there..... Most of us use EMIII or a number of other monitoring programs found in the sticky.... EMIII keeps track of points/etc and will let you know when something is wrong with the client...


Keep on Folding!! For the [H]orde!!

 
The Console is generally prefered, but you can add the flags to your graphical client as well. Just modify the shortcut and add the -advmethods outside of the quotes

 
don't forget "-forceasm" also!


Keep on Folding!! For the [H]orde!!

 
nub to this and hell day at work, but I want to get my 8 machines here going with this
Where do I put in the flags? or a guide? THANK YOU EVERYONE for being nice and dont mind my folding noobi-ness!
 
Another benefit of the console over the graphical client, is that the graphical client has a tendancy of messing with Open GL applications.

Client dosent!

 
Back
Top