mersenneforum.org  

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

Reply
 
Thread Tools
Old 2005-03-19, 20:49   #1
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

61538 Posts
Default memory usage in P-1 stage 1

We all know the memory settings specified in Prime95 apply to P-1, stage2. However, it's not clear from the readme (or anywhere else) that memory usage may exceed those values, and exceed the default/oft-quoted "8MB" value. For example, running P-1 stage 1 of a top-range exponent (M79299821, in my case) uses about 90MB. LL testing seems to use about 45MB for the same exponent. I'm not sure how much memory trial factoring takes up.

I realize this is because of the abnormally-large exponents, but maybe the readme should mention this?
James Heinrich is online now   Reply With Quote
Old 2005-03-19, 21:23   #2
moo
 
moo's Avatar
 
Jul 2004
Nowhere

11001010012 Posts
Default

trial facting takes 1.4 megs of mem when prime95 is in taskbar mode.
moo is offline   Reply With Quote
Old 2005-03-22, 05:24   #3
ewmayer
2ω=0
 
ewmayer's Avatar
 
Sep 2002
Rep├║blica de California

982710 Posts
Default

P-1 stage should need no more memory than an LL test.
ewmayer is offline   Reply With Quote
Old 2005-03-22, 18:26   #4
James Heinrich
 
James Heinrich's Avatar
 
"James Heinrich"
May 2004
ex-Northern Ontario

11×172 Posts
Default

Quote:
Originally Posted by ewmayer
P-1 stage should need no more memory than an LL test.
"should" and "does" don't seem to coincide in this case... P-1,stage1 takes twice the memory that the L-L of the exponent does.
James Heinrich is online now   Reply With Quote
Old 2005-03-22, 19:06   #5
ewmayer
2ω=0
 
ewmayer's Avatar
 
Sep 2002
Rep├║blica de California

982710 Posts
Default

Quote:
Originally Posted by James Heinrich
"should" and "does" don't seem to coincide in this case... P-1,stage1 takes twice the memory that the L-L of the exponent does.
Hm, that almost sounds like stage 1 is using a right-to-left binary powering, which needs 2x the memory of a left-to-right, and roughly 50% more runtime. IMO the only reasons one would ever use right-to-left:

1) The product of all the stage 1 prime powers (needed by the LR powering) is too large to be easily computed/stored - but that needs a very large stage 1 bound;

2) If stage 2 fails to find a factor, one wants to go back and, starting with the original stage 1 residue, do a deeper stage 1 continuation.

George?
ewmayer is offline   Reply With Quote
Old 2005-03-22, 20:05   #6
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

11011111101102 Posts
Default

Maybe we are confusing two terms. Prime95 is concerned with the working set, but perhaps James is reporting the total allocated memory.

P-1 allocates more memory than an LL test because it computes the left-to-right powering exponent. However, only one cache line of this memory will be in prime95's working set during stage 1 - so the two should have the same working set.

Otherwise, I'm at a loss to explain the difference.
Prime95 is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Stage 2 Memory Setting - Again Antonio Software 6 2012-09-04 12:48
Stage 2 Memory Settings gamer30 Software 17 2012-08-23 20:02
Memory usage during P-1 factoring lidocorc Software 2 2008-11-03 02:35
PauseWhileRunning & P-1 memory usage James Heinrich Software 4 2006-01-20 13:55
memory usage in stage 2 of P-1 factoring gckw Software 3 2003-09-07 06:56

All times are UTC. The time now is 01:43.

Thu Nov 26 01:43:52 UTC 2020 up 76 days, 22:54, 3 users, load averages: 1.29, 1.15, 1.31

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, 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.