New Fahcore_78.exe v 1.55?

[primesuspect]

I just deleted my fahcore_78.exe for the hell of it when I downloaded the 4.0 client and now I just got 1.55 dated today, 12/22/2003.

Just a heads up.

[mmonnin]

It prolly wont be forced until awhile. Get it now Terry.;)

[csimon]

I'll have to try it when I finish this wu in about 20 mins

edit:\ got it ...dloaded and purrrin like a kitten!

[edcentric]

What wu were you working on?
I have been reinstalling on a box with 4.0 and in every case it has grabbed the old 78.

[t1rhino]

Only certain servers have the new core. It should be pushed to all the servers soon.

[csimon]

for those who don't get over to the stanford site often here is the announcement with a brief explanation of the v1.55 core:

Hi,
Just wanted to let you know that a new version of FahCore_78 will be out shortly. There's no visible change in it, but an improvement on the science side, with support added for an exciting method which trades some accuracy for a large increase in speed (this won't be noticeable from work unit times, though, as the simulation time covered by a single work unit will be longer in projects using this method). No need to manually delete your current core; whenever the first project assigning units utlizing that method begins, that server will force an update. In relation to units not utilizing that method, the core is unchanged from 1.54.
-Guha

[edcentric]

I need to do some house keeping.
I am running 3.24, 3.25beta, and 4.0 on various machines.
as a result some are using 1.54 and now some 1.55.

There are differences in the types of wu that they get. Time to update everything. Yes the joy of bigger work units.

[Enverex]

Have they fixed the Linux version yet? Before, the SSE version died immediately on lots of systems, and the previous version only works on the first WU (SSE that is).

[csimon]

Enverex had this to say
Have they fixed the Linux version yet? Before, the SSE version died immediately on lots of systems, and the previous version only works on the first WU (SSE that is).
no idea ...check for yourself perhaps.

http://forum.folding-community.org/viewforum.php?f=7

[Straight_Man]

Yes, but Pre2 worked on my box also, so cannot promise my milage and yours will be the same. Final 4.00 is purring here on myLinux box, though. I installed to a user directory, repermed all the files to me as user and enabled read and execute global and write to me as user, run as user in a console, works nice now. Its kicked out a bunch of big WUs since I loaded it, and forcesse works even on a P4. The Barton box has a XP instance on it, and it also has final version of 4.00 and is purring like a kitten. When my new RAM comes, can say the P4 box will fold faster, right now it is slightly limping with PC2100 in it.

One thing that DOES happen, the kerrnel keeps trying to renice it to a higher priority than 19, core also, and both are renice locked at lowest priority Linux will handle.

John.

[Leonardo]

I'd like to try the 4.00 version in my boxes. Will it engage and continue with existing projects if I shut down all instances of folding and drop the new files in the same Folding directories?

[mmonnin]

You should delete the client.cfg file as well since there are new options. I would think it will pick up where it left off but backup anyway to make sure.

[csimon]

Leonardo had this to say
I'd like to try the 4.00 version in my boxes. Will it engage and continue with existing projects if I shut down all instances of folding and drop the new files in the same Folding directories?
Everything should take over just fine Leon ...I transfered all 28 boxes in one day without a problem.
You won't need -forceasm anymore because the client is already optimised to force assembly methods.

Any questions?

recommended p4 flags: Fah4Console.exe -advmethods
recommended xp flags: Fah4Console.exe -advmethods -forcesse

monitor the athlon boxes to be sure it runs stable with the -forcesse flag and it will run very fast ...if unstable remove the flag.