Msieve sourceforge discussion

10metreh · 2009-07-20, 06:37

Quote:

Originally Posted by jasonp

Very few changes from beta 3.

I've decided that continuity of this project requires better resources than I've been giving it, and so there is now an Msieve SourceForge project for hosting the repository and releases. Developers who know the code and want write access are welcome to join. Unfortunately I've only committed revision history starting at v1.41, otherwise the new release would never be ready.

I think the next release will be somewhat delayed; there are big changes I want to make, and now you can watch as I make them :)

Happy factoring,
jasonp

Will you still announce new releases here?

henryzz · 2009-07-20, 07:04

Quote:

Originally Posted by jasonp

Very few changes from beta 3.

I've decided that continuity of this project requires better resources than I've been giving it, and so there is now an Msieve SourceForge project for hosting the repository and releases. Developers who know the code and want write access are welcome to join. Unfortunately I've only committed revision history starting at v1.41, otherwise the new release would never be ready.

I think the next release will be somewhat delayed; there are big changes I want to make, and now you can watch as I make them :)

Happy factoring,
jasonp

does this mean we can download svn versions?

jasonp · 2009-07-20, 12:42

I expect that most day-to-day discussion (and release announcements) will still occur here, although there is a msieve-devel mailing list that you can subscribe to.

And yes, the subversion repository is public, you can download the latest code at any time. I'll still occaisionally make a snapshot tarball and post it there, along with a home-built windows binary. My web page can still host things like beta windows binaries.

May Jeff Gilchrist and I should coordinate offline to add his binaries to the sourceforge page, although perhaps I should just build up a sourceforge web page using his documentation for inspiration. In any case, this belongs in another thread.

Jeff Gilchrist · 2009-07-20, 21:00

Quote:

Originally Posted by jasonp

May Jeff Gilchrist and I should coordinate offline to add his binaries to the sourceforge page, although perhaps I should just build up a sourceforge web page using his documentation for inspiration. In any case, this belongs in another thread.

We can definitely figure something out.

Jeff.

Batalov · 2009-07-31, 01:02

The SF new web design annoys me. A lot of useful features are well-hidden. (It used to be possible to get the .tgz ball without having a svn binary.) Anyway, just to help people how to get the latest version without reading the SF docs:

Code:

mkdir msieve
cd msieve
svn co https://msieve.svn.sourceforge.net/svnroot/msieve/trunk/

mdettweiler · 2009-07-31, 01:59

Quote:

Originally Posted by Batalov

The SF new web design annoys me. A lot of useful features are well-hidden. (It used to be possible to get the .tgz ball without having a svn binary.) Anyway, just to help people how to get the latest version without reading the SF docs:

Code:

mkdir msieve
cd msieve
svn co https://msieve.svn.sourceforge.net/svnroot/msieve/trunk/

I can still get a .tgz without svn, like this:

-Go to http://msieve.svn.sourceforge.net/viewvc/msieve/ to browse the SVN online.
-Click on the link to the "trunk" directory.
-Click "Download GNU tarball" right below the directory listing.

Batalov · 2009-07-31, 02:28

That's great!
They used to have a great button/tab: "More" and once you clicked on it, "SVN" etc tabs showed up... After randomly clicking all over the newly designed pages I found it now, too... It is under "Develop" (!).

Batalov · 2009-08-05, 00:45

The poly selection is performing very well with 1.43.

I've tested (very superficially, ~50 hrs out of many more needed) selection for two numbers:
1. a c171 with gnfs/snfs complexity ratio 0.70 and the best preliminary poly did not compete with the snfs poly, but was only 2x time slower (which is intuitively about right; given much time, the best poly will probably almost match the snfs performance).
2. a c169 with gnfs/snfs complexity ratio 0.68. For this one, the search is still running but the current runner up already sieves 1.5x faster than the snfs poly.

Partially, my exercise was to see if the proverbial 0.7 decision boundary estimate was still right in 2009. (and practical, at the same time)

In short, I'd say, projects of this size (with available significant sieving resources) are quite workable with gnfs/snfs complexity ratio <= 0.68 (maybe 0.69) -- with msieve, not pol51 (haven't tested at this time).

10metreh · 2009-08-05, 06:51

Quote:

Originally Posted by Batalov

The poly selection is performing very well with 1.43.

Is 1.43 out? The sourceforge page only shows 1.42.

Batalov · 2009-08-05, 07:19

1.43 is the current SVN version.
(That's part of the reason why this response is here, in this thread.)

10metreh · 2009-08-05, 10:30

Quote:

Originally Posted by Batalov

1.43 is the current SVN version.
(That's part of the reason why this response is here, in this thread.)

I would have thought the current SVN was 1.42.33 or something like that. When 1.43 does come out, we won't be calling it "1.44", will we?

2009-07-31, 01:02	#5
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3⁶·13 Posts	SVN The SF new web design annoys me. A lot of useful features are well-hidden. (It used to be possible to get the .tgz ball without having a svn binary.) Anyway, just to help people how to get the latest version without reading the SF docs: Code: mkdir msieve cd msieve svn co https://msieve.svn.sourceforge.net/svnroot/msieve/trunk/ Last fiddled with by Batalov on 2009-07-31 at 01:06 Reason: URL gets "beautified"! Added [ code ]

2009-08-05, 00:45	#8
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3⁶×13 Posts	msieve -np The poly selection is performing very well with 1.43. I've tested (very superficially, ~50 hrs out of many more needed) selection for two numbers: 1. a c171 with gnfs/snfs complexity ratio 0.70 and the best preliminary poly did not compete with the snfs poly, but was only 2x time slower (which is intuitively about right; given much time, the best poly will probably almost match the snfs performance). 2. a c169 with gnfs/snfs complexity ratio 0.68. For this one, the search is still running but the current runner up already sieves 1.5x faster than the snfs poly. Partially, my exercise was to see if the proverbial 0.7 decision boundary estimate was still right in 2009. (and practical, at the same time) In short, I'd say, projects of this size (with available significant sieving resources) are quite workable with gnfs/snfs complexity ratio <= 0.68 (maybe 0.69) -- with msieve, not pol51 (haven't tested at this time).

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
msieve on KNL	frmky	Msieve	3	2016-11-06 11:45
SourceForge	Dubslow	Programming	1	2015-05-28 14:39
Msieve on a Mac (Help)	pxp	Msieve	1	2013-02-28 14:56
msieve help	em99010pepe	Msieve	23	2009-09-27 16:13
Msieve 1.10	RedGolpe	Msieve	6	2006-09-07 12:56

2009-07-20, 12:42	#3
jasonp Tribal Bullet Oct 2004 3,541 Posts	I expect that most day-to-day discussion (and release announcements) will still occur here, although there is a msieve-devel mailing list that you can subscribe to. And yes, the subversion repository is public, you can download the latest code at any time. I'll still occaisionally make a snapshot tarball and post it there, along with a home-built windows binary. My web page can still host things like beta windows binaries. May Jeff Gilchrist and I should coordinate offline to add his binaries to the sourceforge page, although perhaps I should just build up a sourceforge web page using his documentation for inspiration. In any case, this belongs in another thread.

2009-07-31, 02:28	#7
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3⁶·13 Posts	That's great! They used to have a great button/tab: "More" and once you clicked on it, "SVN" etc tabs showed up... After randomly clicking all over the newly designed pages I found it now, too... It is under "Develop" (!).

2009-08-05, 07:19	#10
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 22405₈ Posts	1.43 is the current SVN version. (That's part of the reason why this response is here, in this thread.)