mersenneforum.org Should I set MaxStage0Prime in local.txt for mprime?
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

 2021-06-15, 21:36 #1 drkirkby   "David Kirkby" Jan 2021 Althorne, Essex, UK 1AC16 Posts Should I set MaxStage0Prime in local.txt for mprime? I have set Code:  Memory=376832 in local.txt, so P-1 factoring can use 368 GB RAM From the file undoc.txt If doing P-1 with a very large B1, the program uses about 8MB of memory to pre-calculate the product of all small primes below 40 million. The primes above 40 million are processed using a slightly slower method. You can elect to use more memory to move the 40 million threshold higher. In prime.txt enter: MaxStage0Prime=n Should I set MaxStage0Prime too, or is it irreverent with "Memory" set to a large value? If it is not irrelevant, what is a sensible number to set MaxStage0Prime to? Last fiddled with by drkirkby on 2021-06-15 at 22:07
 2021-06-15, 22:02 #2 Viliam Furik   "Viliam Furík" Jul 2018 Martin, Slovakia 10011010112 Posts I don't have the answer, instead, I would like to join with basically the same question. What is the optimal n for 64 GiB of memory? What is the correct n if I want it to use 1 GiB for the primes?
 2021-06-16, 02:42 #3 axn     Jun 2003 2·3·7·112 Posts What is the B1 you are using? If that is less than the 40 million mentioned, you're good. Else set the MaxStage0Prime to the B1 you intend to use. The memory allocated for "stage 0" is independent of the memory to be used for stage 2
2021-06-16, 08:46   #4
drkirkby

"David Kirkby"
Jan 2021
Althorne, Essex, UK

22×107 Posts

Quote:
 Originally Posted by axn What is the B1 you are using? If that is less than the 40 million mentioned, you're good. Else set the MaxStage0Prime to the B1 you intend to use. The memory allocated for "stage 0" is independent of the memory to be used for stage 2
The largest B1 I have used is less than 1 million. I did a PRP test on the 100 million digit M332646233. Kriesel had already done the P-1 factoring on that, with B1=2 million and B2 = 100 million. But I'm not attempting any more 100 million decimal digit numbers - one was enough to convince me that I don't really have the computing resources to make that sensible.

Last fiddled with by drkirkby on 2021-06-16 at 09:20

2021-06-16, 08:57   #5
axn

Jun 2003

2·3·7·112 Posts

Quote:
 Originally Posted by drkirkby Did you type B1 my mistake
Nope. It is applicable for Stage 1 only (Stage 0 is a fast form of Stage 1). Only those of us doing humongous P-1/P+1/ECM on small-ish exponents need to worry about this.

2021-06-16, 09:23   #6
drkirkby

"David Kirkby"
Jan 2021
Althorne, Essex, UK

22·107 Posts

Quote:
 Originally Posted by axn Nope. It is applicable for Stage 1 only (Stage 0 is a fast form of Stage 1). Only those of us doing humongous P-1/P+1/ECM on small-ish exponents need to worry about this.
Sorry, I removed my comments about you possibly meaning B2 when I re-read undoc.txt. I see you quoted me. My apologies. In that case I will not worry about setting MaxStage0Prime

Last fiddled with by drkirkby on 2021-06-16 at 09:24

2021-06-16, 12:06   #7
Viliam Furik

"Viliam Furík"
Jul 2018
Martin, Slovakia

619 Posts

Quote:
 Originally Posted by axn What is the B1 you are using? If that is less than the 40 million mentioned, you're good. Else set the MaxStage0Prime to the B1 you intend to use. The memory allocated for "stage 0" is independent of the memory to be used for stage 2
Thanks, that makes sense...

 Similar Threads Thread Thread Starter Forum Replies Last Post petrw1 Information & Answers 1 2013-10-19 01:35 science_man_88 Aliquot Sequences 185 2011-11-08 12:18 Mattyp101 Conjectures 'R Us 2 2011-02-07 13:53 JuanTutors Software 7 2009-02-01 14:24 Ivan Semenov Teams 1 2004-10-12 21:27

All times are UTC. The time now is 08:24.

Tue Aug 3 08:24:21 UTC 2021 up 11 days, 2:53, 0 users, load averages: 2.32, 2.36, 2.36