mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Software

Reply
 
Thread Tools
Old 2005-11-21, 02:23   #1
drew
 
drew's Avatar
 
Jun 2005

38210 Posts
Default Question about factoring bounds

I just began factoring of a new Mersenne number, and noticed the following on th Prime95 output:

Ignoring suggested B1 value, using B1=395000 from the save file
Ignoring suggested B2 value, using B1=4740000 from the save file

What I had done prior to this was increase the memory usage because it said it was going to put off factoring because of memory availability.

So did the program determine it should use larger bounds because of the available memory, but it got overridden by the formerly determined value? If so, isn't he newer value more appropriate? And why do these bounds get entered in the save file anyway if they're computed so quickly when factoring begins?

Thank in advance for clearing this up for me. I'm a very curious person.

Last fiddled with by drew on 2005-11-21 at 02:33
drew is offline   Reply With Quote
Old 2005-11-22, 00:53   #2
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

22·3·641 Posts
Default

Quote:
Originally Posted by drew
Ignoring suggested B1 value, using B1=395000 from the save file
Ignoring suggested B2 value, using B1
(B2, I presume)
Quote:
=4740000 from the save file

What I had done prior to this was increase the memory usage because it said it was going to put off factoring because of memory availability.

So did the program determine it should use larger bounds because of the available memory,
Yes.
Quote:
but it got overridden by the formerly determined value?
Yes.
Quote:
If so, isn't he newer value more appropriate?
It would be if you were starting fresh from the beginning of stage 2 (in fact, it would have to extend stage 1 to the new higher B1 before restarting stage 2) after you changed the memory assignment, but doing that woud throw away all those teracycles of computation already done in stage 2 using the original bounds! Prime95 doesn't try to calculate whether the benefit of using the larger bounds would offset the waste of computation; it's just assumed that there probably wouldn't be enough extra value to justify extending stage 1 then starting stage 2 over.

(IIRC, if only B2 were increased and B1 were unchanged, then Prime95 could just continue stage 2 from where it left off. But I may be mistaken about that.)
Quote:
And why do these bounds get entered in the save file anyway if they're computed so quickly when factoring begins?
Because of curious persons such as yourself who would screw up stage 2 if Prime95 didn't protect itself from mid-stage-2 resetting of B1, of course! :-)
Quote:
Thank in advance for clearing this up for me. I'm a very curious person.
Good! We need curious folks like you.

(But I'm getting a headache from the bouncing whatzit.)
cheesehead is offline   Reply With Quote
Old 2005-11-22, 01:04   #3
cheesehead
 
cheesehead's Avatar
 
"Richard B. Woods"
Aug 2002
Wisconsin USA

11110000011002 Posts
Default

Hmmm ... I see that my preceding answer assumed that your system had already done part of stage 2 before you changed the memory assignment.

But if it hadn't yet started stage 2, there would be no wastage. ... Probably the code just does what it does anyway in order to protect itself as described above, without taking into comsideration whether there's no wastage.

I could give you a more definite answer if I perused the source code again, but your bouncing whatzit has given me such a headache that I couldn't concentrate if I did that perusal right now.

Oh ... I don't know. GEORGE???

Last fiddled with by cheesehead on 2005-11-22 at 01:05
cheesehead is offline   Reply With Quote
Old 2005-11-22, 02:05   #4
drew
 
drew's Avatar
 
Jun 2005

2×191 Posts
Default

Quote:
Originally Posted by cheesehead
Hmmm ... I see that my preceding answer assumed that your system had already done part of stage 2 before you changed the memory assignment.

But if it hadn't yet started stage 2, there would be no wastage. ... Probably the code just does what it does anyway in order to protect itself as described above, without taking into comsideration whether there's no wastage.

I could give you a more definite answer if I perused the source code again, but your bouncing whatzit has given me such a headache that I couldn't concentrate if I did that perusal right now.

Oh ... I don't know. GEORGE???
Thanks for the response. It looked like it hadn't started stage 2 yet when it did this, which is why I asked. But it makes sense now that I realize it gets saved so it can continue if interrupted. It doesn't matter anymore, though. It didn't find any factors so it is off and LL testing.

...oh, and just for you...
drew is offline   Reply With Quote
Old 2005-11-22, 03:23   #5
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

22·3·641 Posts
Default

The way prime95 does stage 1 makes it possible to increase B1 midway through stage 1, but impossible to decrease B1 midway through stage 1.

Stage 2 / B2 is even more complicated!
Prime95 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
What Bounds to choose, and what are Bounds 144 Information & Answers 5 2017-03-15 13:36
Question on P-1 bounds NBtarheel_33 Math 1 2016-05-09 13:10
Question about parameters and bounds Yamato GMP-ECM 0 2007-01-06 15:29
P-1 bounds calculation question axn Software 2 2006-03-04 16:49
Question about lower bounds for prime forms jasong Miscellaneous Math 3 2006-01-20 22:00

All times are UTC. The time now is 09:49.


Thu Dec 9 09:49:56 UTC 2021 up 139 days, 4:18, 0 users, load averages: 1.77, 1.68, 1.41

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.