mersenneforum.org I think I need to factor some more
 User Name Remember Me? Password
 Register FAQ Search Today's Posts Mark Forums Read

 2019-07-30, 02:49 #1 JuanTutors     Mar 2004 509 Posts I think I need to factor some more I have two 100M digit mersennes that it was suggested should be prefactored deeper than they are so far. Both of them look like this: Code: PRP=HEX,1,2,EXPONENT,-1,77,2 I was told Prime95 no longer deals with the TF of these, so I have some questions. How deeply should I TF them? What program should I use to TF them? Can that program be run concurrently with Prime95? And does Prime95 play nice/cross-share information with the suggested program?
 2019-07-30, 03:45 #2 hansl     Apr 2019 CD16 Posts 100M digit numbers would be in the exponent range of around 332million. mfaktc, using CUDA/Nvidia gpu can TF exponents up to roughly 4.2billion (2^32). If you have AMD gpu, you can use mfakto (not sure if it has exact same limit). If you have a decent GPU, this is a good use for it to run concurrently with prime95, since it uses very little CPU. edit: also I'm not sure if that's true Prime95 not handling TF for that. As far as I knew it works on any exponent up to 1 Billion, but I have not tried it. (Primenet certainly tracks factoring up to 1B at least). Last fiddled with by hansl on 2019-07-30 at 03:49
2019-07-30, 04:32   #3
JuanTutors

Mar 2004

509 Posts

Quote:
 Originally Posted by hansl 100M digit numbers would be in the exponent range of around 332million. mfaktc, using CUDA/Nvidia gpu can TF exponents up to roughly 4.2billion (2^32). If you have AMD gpu, you can use mfakto (not sure if it has exact same limit). If you have a decent GPU, this is a good use for it to run concurrently with prime95, since it uses very little CPU. edit: also I'm not sure if that's true Prime95 not handling TF for that. As far as I knew it works on any exponent up to 1 Billion, but I have not tried it. (Primenet certainly tracks factoring up to 1B at least).
How deeply should I factor though? And are there benchmarks for how long that should take? My exponents are around 332 million.

//Edit re Prime95 not taking care of it, what I was told was Prime95 is 80x slower on that and that it doesn't bother anymore.

Last fiddled with by JuanTutors on 2019-07-30 at 04:34

2019-07-30, 04:41   #4
hansl

Apr 2019

5·41 Posts

Quote:
 Originally Posted by dominicanpapi82 How deeply should I factor though? And are there benchmarks for how long that should take? My exponents are around 332 million.
Probably at least 2^81, based on the highlighted columns here:
https://www.mersenne.ca/status/tf/0/0/3/33000

Last fiddled with by hansl on 2019-07-30 at 04:42

 2019-07-30, 05:38 #5 srow7   Jul 2014 3·17 Posts i can run TF on a gpu upto 81bits for you. it will take about a day for each. you can run LL or PRP while i do TF PM me or post the exponents
2019-07-30, 15:11   #6
JuanTutors

Mar 2004

50910 Posts

Quote:
 Originally Posted by hansl 100M digit numbers would be in the exponent range of around 332million. mfaktc, using CUDA/Nvidia gpu can TF exponents up to roughly 4.2billion (2^32). If you have AMD gpu, you can use mfakto (not sure if it has exact same limit). If you have a decent GPU, this is a good use for it to run concurrently with prime95, since it uses very little CPU. edit: also I'm not sure if that's true Prime95 not handling TF for that. As far as I knew it works on any exponent up to 1 Billion, but I have not tried it. (Primenet certainly tracks factoring up to 1B at least).
My older laptop has the Intel HD Graphics 520 (3.5 years old), and my desktop has the Intel HD Graphics 630 (1.5 years old). Which program would I use for them?

I don't remember what my brand new laptop has but I won't be taxing that one too much ... too expensive (\$3000). I don't want to risk any damage.

Last fiddled with by JuanTutors on 2019-07-30 at 15:11

 2019-07-30, 17:04 #7 hansl     Apr 2019 5×41 Posts You can try mfakto (the "o" is for OpenCL), if those chips have suppor for it. Intel graphics have not been as widely tested with mfakto, compared to Radeon/AMD, but it might work. Just keep in mind that integrated graphics are on the same die as cpu, so thermals may be even more of a concern if running alongside prime95.
2019-07-30, 22:17   #8
JuanTutors

Mar 2004

509 Posts

Quote:
 Originally Posted by hansl You can try mfakto (the "o" is for OpenCL), if those chips have suppor for it. Intel graphics have not been as widely tested with mfakto, compared to Radeon/AMD, but it might work. Just keep in mind that integrated graphics are on the same die as cpu, so thermals may be even more of a concern if running alongside prime95.
Ya I got stuck on step 1.2 trying to install mfakto. What does this mean:

Code:
- Use the VS2010 solution to build the 32-bit or 64-bit binary, or
- use the Makefile to build using MinGW and gcc

 2019-07-30, 22:44 #9 hansl     Apr 2019 5×41 Posts If you don't know what that means (two different ways of compiling from source), I would recommend trying a pre-built release, such as from here: https://download.mersenne.ca/mfakto/mfakto-0.15pre6
2019-08-03, 16:27   #10
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

5,113 Posts

Quote:
 Originally Posted by JuanTutors Ya I got stuck on step 1.2 trying to install mfakto. What does this mean: Code: - Use the VS2010 solution to build the 32-bit or 64-bit binary, or - use the Makefile to build using MinGW and gcc
Asked and answered already, repeatedly. See for example https://www.mersenneforum.org/showth...t=24625&page=5
Not sure why you enjoy wasting other forum users' time.

 2019-08-03, 17:48 #11 ATH Einyen     Dec 2003 Denmark 2·5·313 Posts I factored his exponents to 2^81.

 Thread Tools

 Similar Threads Thread Thread Starter Forum Replies Last Post VictordeHolland Lone Mersenne Hunters 4 2015-03-07 09:29 penguinman007 Factoring 4 2005-08-21 11:19 nfortino Data 6 2004-12-14 19:25 Ivan Semenov Data 2 2004-05-29 14:30 dsouza123 Software 12 2003-08-21 18:38

All times are UTC. The time now is 22:52.

Mon May 10 22:52:35 UTC 2021 up 32 days, 17:33, 1 user, load averages: 2.28, 2.51, 2.69

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.