mersenneforum.org  

Go Back   mersenneforum.org > Factoring Projects > YAFU

Reply
 
Thread Tools
Old 2016-02-04, 21:41   #12
swellman
 
swellman's Avatar
 
Jun 2012

23·359 Posts
Default

No it is not an especially slow time for my computer but definitely on the high side for a SNFS ~240. Lots of scatter in the ETA vs. SNFS curve though.

Your results show that performance in Linux is better than in Windows. I don't disagree, but I eventually left Linux as I got tired of arm wrestling my machines everytime I wanted to use them...
swellman is offline   Reply With Quote
Old 2016-02-24, 09:54   #13
Dan Ee
 
Mar 2013

78 Posts
Default

I am unable to reproduce the failure. Need more details on how to reproduce.

I tried
Code:
D:\Factoring\lasieve4_ivybridge>gnfs-lasieve4I14e.exe -v -a -f 24000000 -c 2000 c189_147_41.txt
gnfs-lasieve4I14e (with asm64): L1_BITS=15, SVN $Revision$
Warning:  lowering FB_bound to 23999999.
FBsize 1506969+0 (deg 6), 2888143+0 (deg 1)
total yield: 3071, q=24002051 (0.09361 sec/rel) ETA 0h00m)
124 Special q, 762 reduction iterations
reports: 34229068->499577->448037->350947->293902->293567
Number of relations with k rational and l algebraic primes for (k,l)=:

Total yield: 3071
0/0 mpqs failures, 2240/2604 vain mpqs
milliseconds total: Sieve 107431 Sched 0 medsched 41680
TD 19527 (Init 77, MPQS 2316) Sieve-Change 118852
TD side 0: init/small/medium/large/search: 2660 217 1398 3785 301
sieve: init/small/medium/large/search: 1248 15476 1039 35173 220
TD side 1: init/small/medium/large/search: 3041 393 1166 3743 383
sieve: init/small/medium/large/search: 1564 16673 942 32846 2250
Content of c189_147_41.txt
Code:
n: 205451388964856467807686090421078666750691138189010020165236543387336128283211048921251478839880150378554118455189058228045898855699277369470472121604743428745579798048420314101101241961387
# 147^41+41^147, difficulty: 237.08, anorm: 6.37e+39, rnorm: -1.95e+45
# scaled difficulty: 237.08, suggest sieving algebraic side
# size = 2.485e-12, alpha = -0.188, combined = 2.415e-13, rroots = 0
type: snfs
size: 237
skew: 14.7100
c6: 1
c0: 10131387
Y1: -509111094534718962173411120845918138561
Y0: 1483273860320763
m: 44008401100288010210378427144625977757864842683924464711360809690583398911228987316128214321844953723667341332495287889981170421945600498539914605816457803581780075658984410989729763013655
alim: 48000000
rlim: 48000000
lpba: 31
lpbr: 31
mfba: 62
mfbr: 62
alambda: 2.5
rlambda: 2.5
I also tried running nfs(the number) in yafu and sieving seems to make progress.
Dan Ee is offline   Reply With Quote
Old 2016-02-24, 14:42   #14
swellman
 
swellman's Avatar
 
Jun 2012

287210 Posts
Default

Are you running on a Windows machine? The bug does not seem to occur in Linux.
swellman is offline   Reply With Quote
Old 2016-02-24, 19:02   #15
Dubslow
Basketry That Evening!
 
Dubslow's Avatar
 
"Bunslow the Bold"
Jun 2011
40<A<43 -89<O<-88

3×2,399 Posts
Default

D:\Factoring\lasieve4_ivybridge>gnfs-lasieve4I14e.exe would certainly seem to indicate Windows
Dubslow is offline   Reply With Quote
Old 2016-02-25, 00:38   #16
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

23A316 Posts
Default

Unsure about the Windoze port, but the original sievers long ago in the past had the childhood disease of sieving quite poorly when some unwritten rules were violated by a poly. (And it did happen with the automatically generated xyyx polys because of the specificity of the construction.) These unwritten rules were:
1. Y1 > 0
2. |Y1| <= |Y0|
These are easily accommodated for -- one by negating the rat'nl poly, and the other, by flipping both polyns on their heads (i.e. swap all poly indices cj, Yi :: i,j for deg side-i). You may want to try that.

EDIT: another, unrelated hunch to try in parallel. Because 147 happens to be 3*72, one can try to lower the skew by leaving 74 out (that is do over-multiply the algebraic poly by 3 but not by 72; leave coeffs as 7^4*x^6 + 3*41^3*y^6, and m=x/y (or y/x as the case may be, I haven't pulled out a napkin to play; this, here, is the napkin ...and its margins are too narrow). Or 3^5*x^6 + 41^3*7^2*y^6 (less likely to be useful)

Last fiddled with by Batalov on 2016-02-25 at 00:50
Batalov is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Suspect result stebbo PrimeNet 23 2017-06-03 11:14
What does my result mean? Unregistered Information & Answers 6 2013-04-16 21:52
Odd result 1997rj7 PrimeNet 2 2009-12-04 08:48
strange GMP-ECM result yoyo GMP-ECM 1 2009-08-08 07:09
New Result R.D. Silverman NFSNET Discussion 1 2005-04-19 23:45

All times are UTC. The time now is 06:40.

Thu Oct 22 06:40:41 UTC 2020 up 42 days, 3:51, 0 users, load averages: 1.01, 1.46, 1.39

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, 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.