Building sr1sieve for RPi - Page 2

wombatman · 2017-03-04, 01:38

Quote:

Originally Posted by ET_

Keep me informed about it :-) I suppose you could get a 2.5x speedup from using the Arm8 code, the FPUs and the 64bits.

Your estimation was pretty good! The exact number jumps around quite a bit, but for the same input, sr1sieve compiled on the Gentoo 64-bit image works at about 11-13M p/sec. It actually started at about 16M, but has dropped a bit. I'll play around with some things like using -q and specifying the L1 and L2 caches to see if that helps any.

On a related note, I'm getting a strange error from the GGNFS sievers. I got the 32-bit versions successfully built on the Gentoo image, and they work fine when run with the rsa100 test file using factLat.pl. If I try to run them with the parameters and command line from the team sieve here, I get an immediate "Aborted" with no other apparent error message. Running the exact same command line on the 32-bit Raspbian-built GGNFS sievers worked flawlessly.

Without an error message, I'm not even sure where to begin.

wombatman · 2017-03-04, 04:04

Got it figured out. The .poly file was very slightly malformed.

henryzz · 2017-03-04, 10:44

At least use the factmsieve.pl script rather than factorlat.pl. The tools used by factlat are only suitable for very small numbers. Either yafu or factmsieve.py are the recommended tools for gnfs currently.

wombatman · 2017-03-04, 16:21

Oh, I'm well aware, and they're what I typically use. I just ran that one to do a quick test. I've since run the lasieve 14e binary by itself without issue.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
sr1sieve with work files?	masser	Conjectures 'R Us	1	2017-12-28 17:43
32-bit of sr1sieve and sr2sieve for Win	pepi37	Software	5	2013-08-09 22:31
Building gcc 4.4.0	CRGreathouse	Software	1	2009-07-07 22:25
sr1sieve and big P	barcode	Software	5	2007-12-27 04:05
sr1sieve 1.2.0	arminius	Riesel Prime Search	14	2007-12-02 22:53

2017-03-04, 04:04	#13
wombatman I moo ablest echo power! May 2013 2⁶×29 Posts	Got it figured out. The .poly file was very slightly malformed.

2017-03-04, 10:44	#14
henryzz Just call me Henry "David" Sep 2007 Liverpool (GMT/BST) 3³·227 Posts	At least use the factmsieve.pl script rather than factorlat.pl. The tools used by factlat are only suitable for very small numbers. Either yafu or factmsieve.py are the recommended tools for gnfs currently.

2017-03-04, 16:21	#15
wombatman I moo ablest echo power! May 2013 2⁶·29 Posts	Oh, I'm well aware, and they're what I typically use. I just ran that one to do a quick test. I've since run the lasieve 14e binary by itself without issue.