mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU Computing (https://www.mersenneforum.org/forumdisplay.php?f=92)
-   -   mfaktc: a CUDA program for Mersenne prefactoring (https://www.mersenneforum.org/showthread.php?t=12827)

TheJudger 2016-10-03 12:10

Extra versions - Wagstaff and/or LessClasses (Windows, CUDA 8.0)
 
Hi again,

[QUOTE=flashjh;444089]I just compiled the 32 and 64 bit versions of lessclass, Wagstaff, and lessclass Wagstaff.

I emailed to Oliver a few minutes ago, I'm sure he'll post them soon ;)[/QUOTE]

Thank you again!

[URL="http://mersenneforum.org/mfaktc/mfaktc-0.21/mfaktc-0.21.win.cuda80.extra-versions.zip"]Here[/URL] are the [I]extra versions[/I] ([B]Wagstaff[/B] and/or [B]LessClasses[/B]), they are an addon the to [URL="http://mersenneforum.org/mfaktc/mfaktc-0.21/mfaktc-0.21.win.cuda80.zip"]basic[/URL] version, just drop the 6 executables into the [I]basic[/I] folder.

Oliver

TheJudger 2016-10-03 12:18

[QUOTE=flashjh;444090]On this note, I'm sad to say that the workhorse 580s are soon obsolete based on CUDA versions -- If they're not already ;)[/QUOTE]

"Sad" but true, in terms of performance per watt (aka energy efficiency) they are sooooooo obsolete. On the other hand we should be happy that we have more advanced hardware today.

Oliver

Dubslow 2016-10-04 08:30

[QUOTE=James Heinrich;444102]Vaguely related: I discovered while writing a script to fetch new worktodo that mfaktc always rewrites all of worktodo.txt and always uses \r\n even if the file had \n line-endings (which mfaktc can read fine).[/QUOTE]

This is a result of the way \n is interpreted by the C compiler. I assume you are using Windows, for gcc/llvm certainly wouldn't use \r\n on and *nix OS.

Basically, all \r\n are translated to \n for the purposes of I/O in C, and vice versa. The mfaktc source code has no place where one may find the "\r" character sequence.

James Heinrich 2016-10-04 14:18

It doesn't really matter where the behaviour originates, it was just one of those gotchas I ran across. I was doing a simple assignment-count check using [i]filesize(worktodo.txt) / strlen('Factor=1234567890,64,67'."\n")[/i] and writing new assignments, and then wondering why my expected line count and actual line count didn't match.

bayanne 2016-10-15 13:10

Gentoo?
 
Will mfaktc run on Gentoo, or is this a CUDA issue?

Dubslow 2016-10-19 12:43

You need CUDA, which means the Nvidia proprietary drivers. I don't know enough to know if said drivers are compatible with Gentoo, but I strongly suspect not.

Mark Rose 2016-10-19 13:38

[QUOTE=Dubslow;445369]You need CUDA, which means the Nvidia proprietary drivers. I don't know enough to know if said drivers are compatible with Gentoo, but I strongly suspect not.[/QUOTE]

They are. My coworker uses the binary drivers with Gentoo.

Dubslow 2016-10-20 04:30

[QUOTE=Mark Rose;445371]They are. My coworker uses the binary drivers with Gentoo.[/QUOTE]

Is that from the official nvidia installation system or by cribbing the binary from some other distro's package?

Mark Rose 2016-10-20 12:56

[QUOTE=Dubslow;445409]Is that from the official nvidia installation system or by cribbing the binary from some other distro's package?[/QUOTE]

I'm not sure. NVidia does provide a script-based installer, so it's probably using that.

storm5510 2016-10-31 20:20

In all the pages of documents I have read, I came across something which indicated mfaktc will be revised so it can communicate with the PrimeNet server. Is this true? If so, any ideas as to when this might happen?

chalsall 2016-10-31 20:27

[QUOTE=storm5510;446080]In all the pages of documents I have read, I came across something which indicated mfaktc will be revised so it can communicate with the PrimeNet server. Is this true? If so, any ideas as to when this might happen?[/QUOTE]

Complete and utter bullshit.

Care to provide evidence to disprove this statement?


All times are UTC. The time now is 23:11.

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