Page 4 of 4

Posted: Wed Dec 05, 2007 1:28 pm
by TheKellieGang
I was running the same work unit as my wife's computer I believe.. p3040_supervillin-03

Posted: Wed Dec 05, 2007 1:40 pm
by toTOW
TheKellieGang wrote:but the difference between the two is bigger than that :)
Do you fold the same project :?: (you can only compare machines when they fold the same project number)

Posted: Wed Dec 05, 2007 2:20 pm
by v00d00
Try using the -forceasm flag to speed up processing (and correct the bad shutdowns you may have force on FAH).

P3040 is a pretty lightweight workunit in the whole scheme of things. If you are running any sort of current hardware it shouldnt be taking much more than a day or two to complete it.

Posted: Wed Dec 05, 2007 8:12 pm
by bruce
v00d00 wrote:Try using the -forceasm flag to speed up processing (and correct the bad shutdowns you may have force on FAH).
No, no, no. ( :P :wink: )

Please don't tell folks that -forceasm speeds up processing. That rumor has been with us much too long and continuing to circulate it is like some of those junk emails about how opening a christmas greeting with a certain name will wipe out everything on your hard-drive and that McAfee and Norton and ... are unable to do anything about it (in a forwarded message with no date but originally circulated about 10 years ago).

The -forceasm flag WILL correct for bad shutdowns (crashes) if you have them but if your machine doesn't crash, it will not do anything for the speed of your processing.

Posted: Wed Dec 05, 2007 8:42 pm
by 7im
-forceasm works around the "using standard loops" issue (not using speed optimizations like SSE) for after a system/client crash and restart. -forceasm enables the client to fold at full speed (not increase folding speed).

This bastardized myth came about from only half-way explaining the usaged of the -forceasm switch, or from misunderstanding what the switch actually did. To a lay-person, it does appear to make the client fold faster (than before when it was just using standard loops), but the switch does NOT make the system fold any faster than it was originally capable of folding (without any crashes, and using SSE, etc.). Point of view / symantical distortion I guess...

Posted: Wed Dec 05, 2007 8:50 pm
by bruce
How about this:

System crashes can slow down folding unless you use -forceasm.

Posted: Fri Dec 07, 2007 2:11 am
by v00d00
Sorry, it was badly phrased.

I wasnt planning to consolidate the speed up myth, although i suspect the OP would benefit from adding the flag.

What i should have done was request the FAHlog.txt i guess.

I really shouldnt type up 5 min posts in 20 secs with over simplification. :oops:

Posted: Fri Dec 14, 2007 11:55 am
by TheKellieGang
Just a thankyou note to those who gave advice on setting up a second console and also setting up the console to act as a service...

Team 95835 doing great now :) thanks again..