RollingStartTime

HiddenWarrior · 2005-12-15, 04:17

I'm useing Prime95 for doing LMH>79.3 (no contacts to server) and noticed that file local.ini is updating when any exponent is going to be written to results.txt. As the time for factoring LMH exponent is ~1 sec (0-52 bits), this file is updating too frequently (the param RollingStartTime). I have a little idea about what for is that one param, and want ot ask:

Is there a way to make Prime95 not to update this file frequently? It will increase the speed of testings as now file IO takes a number of iterations (possibly it can be done only for LMH testers).

Prime95 · 2005-12-15, 05:23

There is no way to change this. AdvancedFactor might be better but will be de-supported in a future release.

You might try factoring to 2^56 or more to decrease I/O overhead as compared to useful crunching time.

moo · 2005-12-15, 05:39

I found a flaw in advanced factor I posted it here a few weeks ago.
http://mersenneforum.org/showthread.php?t=4906
advanced factor causes prime95 to take a week and a half to test a million exponets to 32 bits
well it would take 24 hours to do a million without it.
during the week with advanced factor it was on for 8 hours a day during the week and 24 hours for the weekends.

2005-12-15, 04:17	#1
HiddenWarrior Jun 2003 Russia, Novosibirsk 2×107 Posts	RollingStartTime I'm useing Prime95 for doing LMH>79.3 (no contacts to server) and noticed that file local.ini is updating when any exponent is going to be written to results.txt. As the time for factoring LMH exponent is ~1 sec (0-52 bits), this file is updating too frequently (the param RollingStartTime). I have a little idea about what for is that one param, and want ot ask: Is there a way to make Prime95 not to update this file frequently? It will increase the speed of testings as now file IO takes a number of iterations (possibly it can be done only for LMH testers).

2005-12-15, 05:23	#2
Prime95 P90 years forever! Aug 2002 Yeehaw, FL 19·397 Posts	There is no way to change this. AdvancedFactor might be better but will be de-supported in a future release. You might try factoring to 2^56 or more to decrease I/O overhead as compared to useful crunching time.

2005-12-15, 05:39	#3
moo Jul 2004 Nowhere 809 Posts	I found a flaw in advanced factor I posted it here a few weeks ago. http://mersenneforum.org/showthread.php?t=4906 advanced factor causes prime95 to take a week and a half to test a million exponets to 32 bits well it would take 24 hours to do a million without it. during the week with advanced factor it was on for 8 hours a day during the week and 24 hours for the weekends.