Prime95 version 25.8 - Page 7

Bundu · 2009-01-27, 20:44

As a possible other reason for feature 2)

If the worktodoOnly=1 feature overrode the maxexponents and unreserve days it means that those working on multiple large assignments (for fun) might not get accidental unreserves too? I have noticed that sticking a large LL test on a couple of cores while the others are TF'ing can sometimes trigger an unreserve

garo · 2009-01-27, 22:58

I set UnReserveDays=900 and DaysofWork=1 to make sure I never get any unwanted work from the server. To specifically request no more work from server there is an option NoMoreWork=1.

RTFM - in this case undoc.txt

Bundu · 2009-01-27, 23:25

Lol, I see, maybe someone could delete my dumb posts then :-)

garo · 2009-01-27, 23:32

Edit: Never mind. I figured out that I had two instances of mprime running in the directory and both were reporting results to the server causing this problem.
----------------------------------------------------------------
I've been trying to submit some LMH results and just got this error repeatedly

Code:

[Comm thread Jan 27 23:23] PrimeNet error 11: Server database malfunction
[Comm thread Jan 27 23:23] ar Delete t_gimps_assigned failed: garo GUID: b2f1ce99c66d1b73486c2aed61e86f2e KEY: d94b67732da0387f935e7d5ad31da95d, result_type: 0 exponent: 94443149

followed by:

Code:

[Comm thread Jan 27 23:30] PrimeNet error 13: Server database full or broken
[Comm thread Jan 27 23:30] Insert t_gimps_factors failed: exponent: 94447369, factor: 4714412870811654617
[Comm thread Jan 27 23:30] Visit http://mersenneforum.org for help.
[Comm thread Jan 27 23:30] Will try contacting server again in 3 minutes.

When I terminated mprime, the prime.spl file was gone so I assume I missed submitting some results that went to TF heaven. I'll hunt them in the results.txt but I wanted to report this bug.

Mr. P-1 · 2009-01-28, 12:17

Quote:

Originally Posted by Jeff Gilchrist

Until that feature ever gets implemented you can manually do this in Windows using SetPriority which lets you see and modify priorities at the thread level which Task Manager will not: http://gilchrist.ca/jeff/SetPriority/

I have a Linux box.

garo · 2009-01-28, 12:35

Quote:

Originally Posted by Mr. P-1

I have a Linux box.

Even easier in Linux. Each thread has its own pid. Use ps -efL and there are several mprime threads. You'll need to figure out which one is your target. Presumably by looking at the amount of memory each one is using.

Mr. P-1 · 2009-01-28, 15:00

Quote:

Originally Posted by garo

Even easier in Linux. Each thread has its own pid. Use ps -efL and there are several mprime threads.

Cheers.

Quote:

You'll need to figure out which one is your target. Presumably by looking at the amount of memory each one is using.

How do I find that out?

garo · 2009-01-28, 15:10

top

Mr. P-1 · 2009-01-28, 15:39

Quote:

Originally Posted by garo

top

top -H shows we the two active threads, but it shows the same memory for each. Is there an option to split them out?

starrynte · 2009-01-30, 01:54

is it possible to set prime95 to delay starting worker threads, prime95 not necessarily starting at boot time. (BootDelay only delays that many seconds after boot time, not seconds after starting)

Prime95 · 2009-01-30, 03:24

Quote:

Originally Posted by starrynte

is it possible to set prime95 to delay starting worker threads,

No.

2009-01-27, 23:32	#70
garo Aug 2002 Termonfeckin, IE 2⁴·173 Posts	Edit: Never mind. I figured out that I had two instances of mprime running in the directory and both were reporting results to the server causing this problem. ---------------------------------------------------------------- I've been trying to submit some LMH results and just got this error repeatedly Code: [Comm thread Jan 27 23:23] PrimeNet error 11: Server database malfunction [Comm thread Jan 27 23:23] ar Delete t_gimps_assigned failed: garo GUID: b2f1ce99c66d1b73486c2aed61e86f2e KEY: d94b67732da0387f935e7d5ad31da95d, result_type: 0 exponent: 94443149 followed by: Code: [Comm thread Jan 27 23:30] PrimeNet error 13: Server database full or broken [Comm thread Jan 27 23:30] Insert t_gimps_factors failed: exponent: 94447369, factor: 4714412870811654617 [Comm thread Jan 27 23:30] Visit http://mersenneforum.org for help. [Comm thread Jan 27 23:30] Will try contacting server again in 3 minutes. When I terminated mprime, the prime.spl file was gone so I assume I missed submitting some results that went to TF heaven. I'll hunt them in the results.txt but I wanted to report this bug. Last fiddled with by garo on 2009-01-27 at 23:59

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Prime95 version 27.3	Prime95	Software	148	2012-03-18 19:24
Prime95 version 26.3	Prime95	Software	76	2010-12-11 00:11
Prime95 version 25.5	Prime95	PrimeNet	369	2008-02-26 05:21
Prime95 version 25.4	Prime95	PrimeNet	143	2007-09-24 21:01
When the next prime95 version ?	pacionet	Software	74	2006-12-07 20:30

2009-01-27, 20:44	#67
Bundu Jul 2004 Mid Calder, Scotland 5·37 Posts	As a possible other reason for feature 2) If the worktodoOnly=1 feature overrode the maxexponents and unreserve days it means that those working on multiple large assignments (for fun) might not get accidental unreserves too? I have noticed that sticking a large LL test on a couple of cores while the others are TF'ing can sometimes trigger an unreserve

2009-01-27, 22:58	#68
garo Aug 2002 Termonfeckin, IE 2⁴·173 Posts	I set UnReserveDays=900 and DaysofWork=1 to make sure I never get any unwanted work from the server. To specifically request no more work from server there is an option NoMoreWork=1. RTFM - in this case undoc.txt

2009-01-27, 23:25	#69
Bundu Jul 2004 Mid Calder, Scotland 5·37 Posts	Lol, I see, maybe someone could delete my dumb posts then :-)

2009-01-28, 15:10	#74
garo Aug 2002 Termonfeckin, IE 2⁴×173 Posts	top

2009-01-30, 01:54	#76
starrynte Oct 2008 California 2²×59 Posts	is it possible to set prime95 to delay starting worker threads, prime95 not necessarily starting at boot time. (BootDelay only delays that many seconds after boot time, not seconds after starting)