Folding@Home Weekly Team News - 31st August 2006

RobOC · 4 Sep 2006 at 22:38

SiriusB said:
do the deadlineless WU trick

I take it once setting the deadlineless WU option you then have to wait for the current unit to finish and the client to download its next WU before setting it back again..

SiriusB · 4 Sep 2006 at 22:50

RobOC said:
I take it once setting the deadlineless WU option you then have to wait for the current unit to finish and the client to download its next WU before setting it back again..

Set it to deadlineless WUs while it is still crunching its current WU. When it finishes it will try to get more work and it will fail - the server isn't giving out any deadlineless WUs. So, go back into the options and change it back. Your client will then hook up to a different server and hopefully get you a nice WU.

Make sure you DO NOT have the -advmethods flag set.

SiriusB

RobOC · 4 Sep 2006 at 23:36

ta

altho this means it'll run out of work over night and sit idle for a few hours hehe doh

SiriusB · 4 Sep 2006 at 23:40

RobOC said:
ta altho this means it'll run out of work over night and sit idle for a few hours hehe doh

Rather than losing out on folding time, try just removing the -advmethods flag and see what happens. Worst thing that will happen is you have to wait until later on tomorrow to try the trick. Either that or dont sleep

SiriusB

RobOC · 5 Sep 2006 at 12:32

The trick, she worked!

joeyjojo · 6 Sep 2006 at 17:07

Thought I'd switch advmethods off. Got myself a 600 pointer on the opty straight away, even without changing server. Around 360 ppd, which is nice

.

SiriusB · 6 Sep 2006 at 17:30

From my observations of the last couple of weeks it would seem with -advmethods on you're stuck with 77x projects. -advmethods off yields 21xx projects which are between 392 and 600 points and the 18xx projects of which all have been 153 pointers but there are a couple of larger ones.

SiriusB

RobOC · 7 Sep 2006 at 10:00

wtf

p1487_DPPC_DOPC_CHOL
credit: 1161.00

shadowscotland · 7 Sep 2006 at 11:47

it's not the only one either

p2390_Coiled coil
credit: 1112

deej · 7 Sep 2006 at 13:33

Is that with advmethods off? I'm still running with it active and am getting up to 600 pointers, but nothing that big.

shadowscotland · 7 Sep 2006 at 14:22

I also am running active - but I've never had anything over 600points - but it's on the F@H list so if your lucky it could be yours, guessing it's new out from the beta WU's - ram requirments could be very high though

Beenom · 7 Sep 2006 at 14:51

Can someone help me, im not sure if this is correct or not but...

[10:41:12] Completed 145000 out of 250000 steps (58)
[10:58:24] Writing local files
[10:58:24] Completed 147500 out of 250000 steps (59)
[11:17:44] Writing local files
[11:17:44] Completed 150000 out of 250000 steps (60)
[11:18:02]
[11:18:02] Folding@home Core Shutdown: INTERRUPTED
[11:18:06] CoreStatus = 66 (102)
[11:18:06] + Shutdown requested by user. Exiting.
Folding@Home Client Shutdown.

(I had to restart there, this is a part of my log file from FahMon)

This below is a bit after the previous log data...

[11:19:46] Assembly optimizations on if available.
[11:19:46] Entering M.D.
[11:20:09] (Starting from checkpoint)
[11:20:09] Protein: p2408_Ribo_alanine_sidechain
[11:20:09]
[11:20:10] Writing local files
[11:20:10] Completed 150000 out of 250000 steps (60)
[11:20:13] Extra SSE boost OK.
[11:34:03] CoreStatus = 1 (1)
[11:34:03] Client-core communications error: ERROR 0x1
[11:34:03] Deleting current work unit & continuing...
[11:34:23] Trying to send all finished work units
[11:34:23] + No unsent completed units remaining.
[11:34:23] - Preparing to get new work unit...

Now it has started from the beginning after all that work

What happened there?

Joe42 · 7 Sep 2006 at 16:27

RobOC said:
wtf

p1487_DPPC_DOPC_CHOL
credit: 1161.00

Holy moly. :eek:

I'm stomped. You'll beat my weekly production with two of those!

And whats happened to the team production? Shouldn't borgs be returning or is the end of the holidays having the opposite effect?

shadowscotland · 7 Sep 2006 at 19:38

Beenom - looks like a EUE - early unit end

that happend occationally - especially if maching crashes/OC to high (aka unsatable)/power cut/etc.

Basically anything that happens, that causes Fahcore to stop mid caculation instead of close (did you say yes to 'a programme is not responding - end now' box - that would do it)

Having said all that it only take half a sec to close - and windows will do it for you so it's not normally an issue.

Why did you have to restart?

Jo see this thread
http://forums.overclockers.co.uk/showthread.php?t=17623834