mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Factoring (https://www.mersenneforum.org/forumdisplay.php?f=19)
-   -   Confused w/ GGNFS test-sieving (https://www.mersenneforum.org/showthread.php?t=27115)

Plutie 2021-09-01 04:02

Confused w/ GGNFS test-sieving
 
I'm trying to get used to using GGNFS for small test-sieving runs, but it keeps quitting with a segfault every time I run it. I'm using the latest svn commit, logs are below:


error:

[CODE]../gnfs-lasieve4I14e -r best -f 40000000 -c 2000 -o out -v
gnfs-lasieve4I14e (with asm64): L1_BITS=15, SVN $Revision: 430 $
Warning: lowering FB_bound to 39999999.
FBsize 0+0 (deg 5), 2433653+0 (deg 1)
Segmentation fault (core dumped)[/CODE]This is occurring with many different polynomials, but the one I initially encountered this with is:

[CODE]n: 34166954268223117586694342692657229016949491449884020389380916912810751648547993862665918137895802414087143647172456606010982099550483495609252531569046466515647
Y0: -13603872162482730571699689045336
Y1: 17861405487832127
c0: 697905031491851072220557657635373187641
c1: 540258594997580854184116508876581
c2: -321067102369857480091237224
c3: -24849778917509463073
c4: 3323710111783
c5: 73332
skew: 7208184.563
# lognorm 50.36, E 43.24, alpha -7.13 (proj -1.21), 5 real roots
# MurphyE(Bf=1.000e+07,Bg=5.000e+06,area=1.000e+16)=1.265e-12
rlim: 67000000
alim: 67000000
lpbr: 30
lpba: 30
mfbr: 60
mfba: 60
rlambda: 2.6
alambda: 2.6
type: gnfs
lss: 0
[/CODE]


Any help would really be appreciated, thank you!

VBCurtis 2021-09-01 04:46

Do you have evidence that your compiled GGNFS binaries work? That is, do they function when called by factmsieve.py or yafu?

If not, acquire working binaries or run one of those scripts to determine if your binaries work. I've never successfully compiled GGNFS, but I haven't tried very hard; there have been almost no changes to the code in ten years.

Plutie 2021-09-01 05:25

[QUOTE=VBCurtis;586966]Do you have evidence that your compiled GGNFS binaries work? That is, do they function when called by factmsieve.py or yafu?

If not, acquire working binaries or run one of those scripts to determine if your binaries work. I've never successfully compiled GGNFS, but I haven't tried very hard; there have been almost no changes to the code in ten years.[/QUOTE]

YAFU's GGNFS calls work perfectly. [CODE]
elapsed time: 20.1866 seconds (101 second deadline); poly select done
nfs: commencing algebraic side lattice sieving over range: 900000 - 901667
nfs: commencing algebraic side lattice sieving over range: 918337 - 920004
nfs: commencing algebraic side lattice sieving over range: 901667 - 903334
nfs: commencing algebraic side lattice sieving over range: 905001 - 906668
nfs: commencing algebraic side lattice sieving over range: 908335 - 910002
nfs: commencing algebraic side lattice sieving over range: 903334 - 905001
nfs: commencing algebraic side lattice sieving over range: 906668 - 908335
nfs: commencing algebraic side lattice sieving over range: 910002 - 911669
nfs: commencing algebraic side lattice sieving over range: 911669 - 913336
nfs: commencing algebraic side lattice sieving over range: 913336 - 915003
nfs: commencing algebraic side lattice sieving over range: 915003 - 916670
nfs: commencing algebraic side lattice sieving over range: 916670 - 918337
total yield: 20462, q=903337 (0.00065 sec/rel)
total yield: 23162, q=915007 (0.00063 sec/rel)
total yield: 23559, q=905011 (0.00065 sec/rel)
total yield: 24649, q=913337 (0.00065 sec/rel)
total yield: 25268, q=906673 (0.00065 sec/rel)
total yield: 26434, q=918341 (0.00063 sec/rel)
total yield: 26316, q=920011 (0.00064 sec/rel)
total yield: 26222, q=911671 (0.00065 sec/rel)
total yield: 27597, q=901679 (0.00062 sec/rel)
total yield: 27775, q=910003 (0.00064 sec/rel)
total yield: 29296, q=908353 (0.00062 sec/rel)
total yield: 30366, q=916679 (0.00062 sec/rel)[/CODE]

unconnected 2021-09-01 06:23

You're sieving on rational side while the posted poly is for algebraic side. Did you try -a instead of -r option?

henryzz 2021-09-01 09:28

[QUOTE=unconnected;586970]You're sieving on rational side while the posted poly is for algebraic side. Did you try -a instead of -r option?[/QUOTE]
That will affect efficiency not cause it to crash.



Two ideas:

Can the ggnfs siever understand lss? At best it will ignore it I think.

We have had issues in the past with compilers making it so that the -r or -a option must be the last option.

Plutie 2021-09-01 12:27

[QUOTE=henryzz;586973]That will affect efficiency not cause it to crash.



Two ideas:

Can the ggnfs siever understand lss? At best it will ignore it I think.

We have had issues in the past with compilers making it so that the -r or -a option must be the last option.[/QUOTE]

Tried both of these, no success.

kruoli 2021-09-01 13:58

So both ideas produced an identical segmentation fault?

Edit: Can you build it with -g and run it in GDB? Do you have experience with that?

Plutie 2021-09-01 14:22

[QUOTE=kruoli;586989]So both ideas produced an identical segmentation fault?

Edit: Can you build it with -g and run it in GDB? Do you have experience with that?[/QUOTE]

[CODE]~/Math/ggnfs/src/experimental/lasieve4_64$ gdb --args gnfs-lasieve4I14e -r ../../../bin/testing/best -o out -f 40000000 -c 2000
GNU gdb (Ubuntu 9.2-0ubuntu1~20.04) 9.2
Copyright (C) 2020 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from gnfs-lasieve4I14e...
(gdb) run
Starting program: /home/uwu/Math/ggnfs/src/experimental/lasieve4_64/gnfs-lasieve4I14e -r ../../../bin/testing/best -o out -f 40000000 -c 2000

Program received signal SIGSEGV, Segmentation fault.
0x0000000000427e30 in slinie_fbi_loop ()
(gdb) bt
#0 0x0000000000427e30 in slinie_fbi_loop ()
#1 0x00000000005d5760 in smallsieve_aux2 ()
#2 0x0000000000000005 in ?? ()
#3 0x0000000000406c76 in main (argc=<optimized out>, argv=<optimized out>)
at gnfs-lasieve4e.c:2085[/CODE]

I should have probably mentioned this earlier, but I'm using EdH's patches for GGNFS from [url]https://www.mersenneforum.org/showthread.php?t=23081[/url].

EdH 2021-09-01 14:38

[QUOTE=Plutie;586991]. . .
I should have probably mentioned this earlier, but I'm using EdH's patches for GGNFS from [URL]https://www.mersenneforum.org/showthread.php?t=23081[/URL].[/QUOTE]
I have not tested these recently.* I will try to do that later today. (BTW, thank you for mentioning this. I like to know if any of my "How I ..." threads don't work. They can quite easily become out of date.)

I can't take credit as the source of the patches. I got them from someone else on the forum, but I can't remember who that may have been.

* I've just been using bsquared binaries that are available in the attachments to that first post.

Plutie 2021-09-01 14:58

Changed binaries, it's working perfectly now. Thank you!

EdH 2021-09-01 15:51

[QUOTE=EdH;586993]I have not tested these recently.* I will try to do that later today. (BTW, thank you for mentioning this. I like to know if any of my "How I ..." threads don't work. They can quite easily become out of date.)
. . .[/QUOTE]I used my "How I . . ." with a Xeon/Ubuntu machine and all went well with the above supplied poly:[code]$ ../gnfs-lasieve4I14e -r testPoly -f 40000000 -c 2000 -o testRels -v
gnfs-lasieve4I14e (with asm64): L1_BITS=15, SVN $Revision$
Warning: lowering FB_bound to 39999999.
FBsize 3951781+0 (deg 5), 2433653+0 (deg 1)
total yield: 2205, q=40002023 (0.22094 sec/rel)
118 Special q, 289 reduction iterations
reports: 5785443->2476959->2088570->890333->373860->275017
Number of relations with k rational and l algebraic primes for (k,l)=:

Total yield: 2205
0/0 mpqs failures, 1633/711 vain mpqs
milliseconds total: Sieve 163389 Sched 0 medsched 71955
TD 28275 (Init 732, MPQS 1298) Sieve-Change 223578
TD side 0: init/small/medium/large/search: 3026 2316 1769 5295 1933
sieve: init/small/medium/large/search: 2257 34308 1797 48850 2489
TD side 1: init/small/medium/large/search: 3363 891 1734 4778 809
sieve: init/small/medium/large/search: 1523 20163 1794 49621 586[/code]Since this worked and therefore I couldn't duplicate the original trouble, and since you have solved your trouble by using the alternate binaries, I'll not pursue this further for now. I would be interested if you revisit this in the future.

Thanks and glad you're working properly.


All times are UTC. The time now is 01:40.

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