PDA

View Full Version : Client selection



vaughan
04-23-2004, 03:17 AM
Is there a way to select only a certain type of clients in FAH4?

Bionic_Redneck
04-23-2004, 05:29 AM
try the -advmethods flag if your use the text client or if your if you have the gui it can be selected by right clicking on icon in taskbar. It's not set in stone but should get gromac wu with this method but if you want tinkers you should just leave the advance methods alone.

vaughan
04-23-2004, 06:22 AM
Are tinkers better suited to XPs and gromacs for P4s?

Bionic_Redneck
04-23-2004, 06:50 AM
gromacs are good for both cpus and tinkers are good for any athlon or p3 but are very slow with p4 and now they have another gromacs core called core_79 which uses SSE2 so XP's can't use them and run very fast from what I hear on p4's.

SB2
04-23-2004, 11:14 AM
So this "new" scoring system for "tinker" wu's makes this a viable project for AMD syystems again? And has the new core disabled the -forcesse switich for AMD's? As I remember the SSE code had no problems on the earlier XP's (1700+ - 1800+}.

Bionic_Redneck
04-24-2004, 04:57 PM
no you can still use -forcesse I have never had any issues myself with any the xp's I have had some are older paliminos. The problem with sse and folding@home has been worked on by AMD/Pande group and should have been fixed long ago

SB2
04-24-2004, 06:02 PM
Was just looking over Short-Media's (http://short-media.com/forum/showthread.php?t=12518) forum, and all indications are the new point system for tinker workunits is having a big impact on the point production of AMD's. This should prove to be a big boost to the team if we can get a few more people folding on a regular basis.

Also note (i've been testing on 2 P4's (http://24.174.34.241:8080/Foldingstats.html) )
the difference in the _78 and _79 cores, and it appears that the _79 core is 3 to 4 times as fast as the _78 core.


no you can still use -forcesse I have never had any issues myself with any the xp's I have had some are older paliminos. The problem with sse and folding@home has been worked on by AMD/Pande group and should have been fixed long ago

That is good to know. I'm going to test the (new) tinker scoring comparing it to the gromacs with -forcesse switich on a couple of XP's. to see if indeed the tinkers will now out produce the gromacs workunits.

Bionic_Redneck
04-24-2004, 06:30 PM
I thought about this just now, if your running linux on any of your computers with tinkers it's slower than with windows but if you have windows client you can run it with wine just don't go with gui as if you open window to show grapics it locks the whole computer up

Bionic_Redneck
04-26-2004, 01:55 PM
it appears that you still get gromacs without using advmethods so I guess they are now main stream and there is no way of choosing tinkers only what i'm going to do is make anther thread showing what point averages we would get from different WU's since the node are all using same hardware should work out good.