Package Details: mprime 307b9-2

Git Clone URL: https://aur.archlinux.org/mprime.git (read-only, click to copy)
Package Base: mprime
Description: A GIMPS, distributed computing project client, dedicated to finding Mersenne primes.
Upstream URL: http://www.mersenne.org
Licenses: custom
Conflicts: mprime-bin
Submitter: graysky
Maintainer: graysky
Last Packager: graysky
Votes: 71
Popularity: 0.32
First Submitted: 2013-12-11 21:18 (UTC)
Last Updated: 2022-05-05 18:19 (UTC)

Latest Comments

Shished commented on 2022-05-17 12:03 (UTC)

When app is launched dmesg shows this message:

process 'arch/usr/bin/mprime' started with executable stack

What is that?

Iiridayn commented on 2020-08-14 01:05 (UTC)

It might be useful to note clearly that this source version of mprime can not communicate with the work server. This is a deliberate choice, to prevent a modified client from spamming the server with fake results. Thus, work files will need to be downloaded/uploaded manually, if using it to find Mersenne Primes instead of stress testing.

graysky commented on 2020-03-23 12:28 (UTC) (edited on 2020-03-23 12:29 (UTC) by graysky)

error libhwloc.so.5 :(

It builds and runs for me.

tdolega commented on 2020-03-23 11:29 (UTC)

error libhwloc.so.5 :(

graysky commented on 2020-03-19 08:34 (UTC)

@aqua - You have to rebuild it because the underlying libs changes.

graysky commented on 2020-03-19 08:33 (UTC) (edited on 2020-03-21 10:56 (UTC) by graysky)

@yochananmarqos - not out-of-date until they post source for 298b7 ... https://mersenneforum.org/showthread.php?p=540358

emiliobadillio commented on 2020-03-18 14:41 (UTC)

@aqua123 I got that same error when installing from pacman, but if you build it from cloning the repository, it works. New issue when running a stress test, it will kill the process after running once on each cpu.

[Main thread Mar 18 10:38] Starting workers.
[Worker #1 Mar 18 10:38] Worker starting
[Worker #2 Mar 18 10:38] Worker starting
[Worker #3 Mar 18 10:38] Worker starting
[Worker #4 Mar 18 10:38] Worker starting
[Worker #1 Mar 18 10:38] Beginning a continuous self-test on your computer.
[Worker #1 Mar 18 10:38] Please read stress.txt.  Hit ^C to end this test.
[Worker #3 Mar 18 10:38] Beginning a continuous self-test on your computer.
[Worker #3 Mar 18 10:38] Please read stress.txt.  Hit ^C to end this test.
[Worker #4 Mar 18 10:38] Beginning a continuous self-test on your computer.
[Worker #4 Mar 18 10:38] Please read stress.txt.  Hit ^C to end this test.
[Worker #2 Mar 18 10:38] Beginning a continuous self-test on your computer.
[Worker #2 Mar 18 10:38] Please read stress.txt.  Hit ^C to end this test.
[Worker #3 Mar 18 10:38] Test 1, 52000 Lucas-Lehmer iterations of M5705025 using AVX FFT length 288K, Pass1=384, Pass2=768, clm=2.
[Worker #4 Mar 18 10:38] Test 1, 52000 Lucas-Lehmer iterations of M5705025 using AVX FFT length 288K, Pass1=384, Pass2=768, clm=2.
[Worker #1 Mar 18 10:38] Test 1, 52000 Lucas-Lehmer iterations of M5705025 using AVX FFT length 288K, Pass1=384, Pass2=768, clm=2.
[Worker #2 Mar 18 10:38] Test 1, 52000 Lucas-Lehmer iterations of M5705025 using AVX FFT length 288K, Pass1=384, Pass2=768, clm=2.
Killed

aqua123 commented on 2020-03-18 14:02 (UTC)

I'm getting

mprime: error while loading shared libraries: libhwloc.so.5: cannot open shared object file: No such file or directory

when I try to run mprime. In /usr/lib/ I find libhwloc.so, libhwloc.so.15, and libhwloc.so.15.1.0. I've tried to symlink libhwloc.so.5 to one of these but I keep getting segfaults when the stress tests start so I think I'm doing something wrong - is this an issue that anyone else is facing?

gee commented on 2019-04-14 01:28 (UTC)

Thank you for the quick turnaround!

graysky commented on 2019-04-13 17:21 (UTC)

Oops, I didn't mean to push the updated version. For some reason, hwloc in in [extra] is outdated. Once it goes to 2.0.3 this will work. Until then I will push 294b7-4 back.

gee commented on 2019-04-13 17:08 (UTC)

depends = hwloc>2.0.3

Is it a good idea to depend on something not found even in aur?

CyberShadow commented on 2018-04-12 22:42 (UTC)

I think make will only pull things from the environment if the variables are set with := , not = . Is that why you re-added the makefile patch?

In any case, what I meant was: make CFLAGS="..." CXXFLAGS="..." etc.

Also, not sure why the variables were added to the package function - the make invocation happens in the build function :)

I guess I should try submitting a tested patch rather than vague suggestions.

graysky commented on 2018-04-12 18:06 (UTC) (edited on 2018-04-12 18:10 (UTC) by graysky)

@TX1683 - Fixed in 294b7-3 @arcnmx - Odd... built fine for me last night but not now. Fuck it. 294b7-4.

arcnmx commented on 2018-04-12 17:02 (UTC)

In addition to that error, using --skipinteg fails later on when linking (it seems to try to statically link curl and hwloc, while only .so is present?)

TX1683 commented on 2018-04-12 08:58 (UTC)

There is problem in latest PKGBUILD.

==> ERROR: Integrity checks (sha256) differ in size from the source array. ==> ERROR: Makepkg was unable to build mprime.

graysky commented on 2018-04-11 21:27 (UTC)

@CyberShadow - Guess not but it really doesn't matter.

CyberShadow commented on 2018-04-11 19:22 (UTC)

Sorry, is unfuck_makefile.patch really needed? The patch just seems to adjust some variable definitions, but you can override make variables by placing them on its command-line.

graysky commented on 2017-10-18 19:22 (UTC)

@sergio - Thanks for that. Will test and update.

qrwteyrutiyoup commented on 2017-10-18 15:06 (UTC)

@graysky: here's a diff against revision "bc91d9" ("Update to 291-5") that updates it to 293-1: https://gist.github.com/qrwteyrutiyoup/4682d87516b1199fc424212ed8b1674f

graysky commented on 2017-10-09 19:24 (UTC)

Modification to the makefile is needed for the 293 release but I am unable to troubleshoot now. Anyone care to investigate?

graysky commented on 2017-05-13 10:41 (UTC)

@cinna - Upstream is to blame for this... it is really annoying.

cinna commented on 2017-05-13 08:22 (UTC)

Is there a reason why the checksum keeps changing for the same version? Checksum is now "b2af302d423765e1f053731f4c0dac6983a9598eb1b16fcf6159491dbb1c3972"

jugs commented on 2017-04-08 18:09 (UTC)

Checksum changed again e5aada84256863882823d08e87495118c4b19e0fa03f3105da4c2fbcd26c6045 p95v291.source.zip

graysky commented on 2017-03-21 18:32 (UTC)

Not sure why the checksum changed from when I built it yesterday. 291-3 fixes that.

gavinhungry commented on 2017-03-21 16:33 (UTC)

@kerberizer, you're not mistaken. Also, url and source may use SSL for mersenne.org.

kerberizer commented on 2017-03-21 12:28 (UTC)

I might be mistaken, but it seems that the sha256sum of p95v291.source.zip is "a065c0b16de8c6dbfdcf7a8e1e07c9d77cc7dbbdd882ca6f03a62a769fc4a76a".

graysky commented on 2017-03-20 11:19 (UTC)

Upstream needs to update their website (https://www.mersenne.org/download/) which is advertising 2810 as the current... as we sure that 29.1 is considered stable?

graysky commented on 2016-03-10 21:03 (UTC)

@coruun - Good question! Fixed in 287-3, thanks for pointing it out.

coruun commented on 2016-03-10 08:13 (UTC)

Why do you patch gwnum/makefile, while you are actually using gwnum/make64 for compilation?

graysky commented on 2016-02-18 19:10 (UTC)

@ase - All members of base-devel are assumed to be installed when using the AUR.[1] 1. https://wiki.archlinux.org/index.php/Arch_User_Repository#Prerequisites

ase1590 commented on 2016-02-18 15:12 (UTC)

Add "patch" as a dependency please.

graysky commented on 2015-09-21 06:29 (UTC)

@firekage - Works for me without any switches and no hung processes. The mprime-bin is packaged with some docs but no formal manpage.

firekage commented on 2015-09-21 03:27 (UTC)

I have one problem, maybe somebody will help me. I installed mprime using aur. For the first time it worked, could do some test and could choose which test to do. After finish i hit the exit in terminal, and after that, as soon as i start mprime again i see that command prompt is being busy but nothing happens, i don't have list with test and choices given at all. What's wrong? I type "mprime" in terminal and it either stops command prompt or give an message: "another mprime is already run" where in fact, there is no a such process at all! Edit: there is no manpage for mprime. On mprime main website i saw that under linux it should be run wtich a switch: mprime -m.

graysky commented on 2015-05-28 19:18 (UTC)

Bump to v287-1

graysky commented on 2015-05-28 15:10 (UTC)

I will update to 286 once upstream fixes their invalid zip archive.

graysky commented on 2014-10-09 04:39 (UTC)

Ha! Got it working again. Enjoy!

graysky commented on 2014-08-13 19:22 (UTC)

Disowned. Maybe someone else with more time and know-how can update.

graysky commented on 2014-06-04 21:34 (UTC)

Gonna need some help getting 285 to build. Please submit changes to the PKGBUILD here.

graysky commented on 2014-02-08 13:44 (UTC)

Builds fine for me as-is. Suggest you delete your source zip and try again.

0x47 commented on 2014-02-07 20:53 (UTC)

sha256sums fails on this package...

graysky commented on 2013-12-21 11:23 (UTC)

Hi Xyne, thanks for the clean up. Actually i686 is supported, I just can't get it to build. Perhaps you can?

Xyne commented on 2013-12-21 01:34 (UTC)

Hi, I thought the PKGBUILD was a bit messy so I cleaned it up. If i686 is no longer supported (it was commented in your PKGBUILD), then please remove all of the conditional statements. You can find the package source archive here: http://xyne.archlinux.ca/tmp/pkgbuilds/ Regards, Xyne