Msieve 1.42 feedback - Page 4

10metreh · 2009-07-08, 07:29

Quote:

Originally Posted by jrk

Oops I found a bug in beta 2.

<snip>

Can someone compile some binaries with this change?

Shaopu Lin · 2009-07-08, 10:27

When I cracked rsa100 by 1.42-beta2, it crashed due to segmentation fault.

Code:

math@linux-0rdr:~/Desktop/application/nfs/ggnfs/gnfs> gdb msieve
GNU gdb (GDB; openSUSE 11.1) 6.8.50.20081120-cvs
Copyright (C) 2008 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-suse-linux".
For bug reporting instructions, please see:
<http://bugs.opensuse.org/>...
(gdb) run  -s gnfs.dat -l ggnfs.log -i gnfs.ini -v -nf gnfs.fb -t 2 -nc1
Starting program: /usr/local/bin/msieve -s gnfs.dat -l ggnfs.log -i gnfs.ini -v -nf gnfs.fb -t 2 -nc1
[Thread debugging using libthread_db enabled]


Msieve v. 1.42
Wed Jul  8 18:21:08 2009
random seeds: 77e196ce 85d7a6dd
factoring 8152076978298033032614464108763248275638533543184212928005190196721776176660059482366247964158442259 (100 digits)
searching for 15-digit factors
[New Thread 0x7ffff6fbe950 (LWP 7890)]
commencing number field sieve (100-digit input)
R0: -1051744772620805140007329
R1:  137992068071417
A0:  3371890176974270511868545840
A1: -10886188273030830462761
A2: -26456360310836702
A3:  17842052158
A4:  6660
skew 1256178.72, size 1.013689e-13, alpha -5.324659, combined = 1.211124e-08

commencing relation filtering
estimated available RAM is 925.9 MB
commencing duplicate removal, pass 1
found 153747 hash collisions in 2488365 relations
commencing duplicate removal, pass 2
found 77110 duplicates and 2411255 unique relations
memory use: 8.2 MB
reading ideals above 100000
commencing singleton removal, initial pass
memory use: 74.6 MB
reading all ideals from disk
memory use: 74.0 MB
commencing in-memory singleton removal
begin with 2411255 relations and 3475561 unique ideals
reduce to 366 relations and 0 ideals in 9 passes
max relations containing the same ideal: 0

Program received signal SIGSEGV, Segmentation fault.
0x0000000000438890 in filter_merge_init ()
Current language:  auto; currently asm
(gdb) bt full
#0  0x0000000000438890 in filter_merge_init ()
No locals.
#1  0x00000000004368eb in filter_make_relsets ()
No locals.
#2  0x0000000000424f62 in nfs_filter_relations ()
No locals.
#3  0x000000000041af8a in factor_gnfs ()
No locals.
#4  0x00000000004067a9 in msieve_run ()
No locals.
#5  0x0000000000404557 in factor_integer ()
No locals.
#6  0x0000000000404f91 in main ()
No locals.
(gdb)

My msieve is compiled with ECM=1 on 64-bit OpenSUSE 11.1.

jasonp · 2009-07-08, 12:42

jrk, thanks for finding my cut-n-paste error. That's been fixed along with a few other small things, and I've reposted the new source and binary at the same URLs. Shaopu, this was also the cause of your crash.

The crash during singleton removal is a little odd. If the relations file is not too large can you post it on rapidshare? If not, please at least PM me the .fb file and log file.

Shaopu Lin · 2009-07-08, 15:42

Quote:

Originally Posted by jasonp

The crash during singleton removal is a little odd. If the relations file is not too large can you post it on rapidshare? If not, please at least PM me the .fb file and log file.

The relations and other files are available from http://rapidshare.com/files/253443873/gnfs.tar.bz2.html.

Batalov · 2009-07-08, 15:49

I had exit 11, when filtering (new style, immediately in memory) on a very small gnfs, on linux x86_64. The available memory was determined fairly correctly by get_ram_size(). Then some filtering and exit 11.

jasonp · 2009-07-08, 16:03

Sorry, I was referring to jrk's second bug report, with the problems during singleton removal. Shaopu, do you have the same problem now that the source has been updated?

jrk · 2009-07-08, 17:34

Quote:

Originally Posted by jasonp

The crash during singleton removal is a little odd. If the relations file is not too large can you post it on rapidshare? If not, please at least PM me the .fb file and log file.

Uploaded: msieve-singleton-removal-crash.tar.bz2 (26.4 MB)

The crash can be triggered by running

Code:

msieve -v -nc -mb 20 1827236742412001149081161888382165147842309369237104162298150177723605901079739581470637

Note the -mb 20. While re-running msieve to try to find the first bug, msieve would sometimes estimate the available memory low, and crash in singleton removal instead.

gdb shows that on common/filter/singleton.c:411 ideal=4294967295 (-1) which results in an out-of-bounds array access on freqtable[ideal].

Shaopu Lin · 2009-07-08, 18:50

Quote:

Originally Posted by jasonp

Sorry, I was referring to jrk's second bug report, with the problems during singleton removal. Shaopu, do you have the same problem now that the source has been updated?

I tested the binary compiled from the new code, the same problem no longer occured.

Code:

Thu Jul  9 01:33:12 2009  
Thu Jul  9 01:33:12 2009  
Thu Jul  9 01:33:12 2009  Msieve v. 1.42
Thu Jul  9 01:33:12 2009  random seeds: 7a48fd28 3c80b71c
Thu Jul  9 01:33:12 2009  factoring 8152076978298033032614464108763248275638533543184212928005190196721776176660059482366247964158442259 (100 digits)
Thu Jul  9 01:33:13 2009  searching for 15-digit factors
Thu Jul  9 01:33:15 2009  commencing number field sieve (100-digit input)
Thu Jul  9 01:33:15 2009  R0: -1051744772620805140007329
Thu Jul  9 01:33:15 2009  R1:  137992068071417
Thu Jul  9 01:33:15 2009  A0:  3371890176974270511868545840
Thu Jul  9 01:33:15 2009  A1: -10886188273030830462761
Thu Jul  9 01:33:15 2009  A2: -26456360310836702
Thu Jul  9 01:33:15 2009  A3:  17842052158
Thu Jul  9 01:33:15 2009  A4:  6660
Thu Jul  9 01:33:15 2009  skew 1256178.72, size 1.013689e-13, alpha -5.324659, combined = 1.211124e-08
Thu Jul  9 01:33:15 2009  
Thu Jul  9 01:33:15 2009  commencing relation filtering
Thu Jul  9 01:33:15 2009  estimated available RAM is 747.5 MB
Thu Jul  9 01:33:15 2009  commencing duplicate removal, pass 1
Thu Jul  9 01:33:28 2009  found 145935 hash collisions in 2363660 relations
Thu Jul  9 01:33:33 2009  added 124705 free relations
Thu Jul  9 01:33:33 2009  commencing duplicate removal, pass 2
Thu Jul  9 01:33:34 2009  found 77110 duplicates and 2411255 unique relations
Thu Jul  9 01:33:34 2009  memory use: 8.2 MB
Thu Jul  9 01:33:34 2009  reading ideals above 100000
Thu Jul  9 01:33:34 2009  commencing singleton removal, initial pass
Thu Jul  9 01:33:53 2009  memory use: 74.6 MB
Thu Jul  9 01:33:53 2009  reading all ideals from disk
Thu Jul  9 01:33:53 2009  memory use: 74.0 MB
Thu Jul  9 01:33:53 2009  commencing in-memory singleton removal
Thu Jul  9 01:33:53 2009  begin with 2411255 relations and 3475561 unique ideals
Thu Jul  9 01:33:53 2009  reduce to 366 relations and 0 ideals in 9 passes
Thu Jul  9 01:33:53 2009  max relations containing the same ideal: 0
Thu Jul  9 01:33:53 2009  filtering wants 65664 more relations
Thu Jul  9 01:33:53 2009  elapsed time 00:00:41

jasonp · 2009-07-09, 18:27

jrk, I downloaded your relations but the crash doesn't happen with the latest code. Does it work for you now?

jrk · 2009-07-09, 23:27

Quote:

Originally Posted by jasonp

jrk, I downloaded your relations but the crash doesn't happen with the latest code. Does it work for you now?

No it still crashes.

Try this one: msieve-singleton-removal-crash2.tar.bz2

Again using:

Code:

msieve -v -nc -mb 20 498164440375209611062493528103328938326299406869997774494699343920871291837701968386330622179690071799

Interestingly, if I trim the relations down to 1.4M, it does not crash. Although it needed more relations, not less.

jasonp · 2009-07-10, 00:33

Grr, still no crash when run locally. Also no crash on my 64-bit system. What platform and compiler version are you using?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Msieve 1.53 feedback	xilman	Msieve	149	2018-11-12 06:37
Msieve 1.50 feedback	firejuggler	Msieve	99	2013-02-17 11:53
Msieve v1.48 feedback	Jeff Gilchrist	Msieve	48	2011-06-10 18:18
Msieve 1.43 feedback	Jeff Gilchrist	Msieve	47	2009-11-24 15:53
Msieve 1.41 Feedback	Batalov	Msieve	130	2009-06-09 16:01

2009-07-08, 12:42	#36
jasonp Tribal Bullet Oct 2004 3,541 Posts	jrk, thanks for finding my cut-n-paste error. That's been fixed along with a few other small things, and I've reposted the new source and binary at the same URLs. Shaopu, this was also the cause of your crash. The crash during singleton removal is a little odd. If the relations file is not too large can you post it on rapidshare? If not, please at least PM me the .fb file and log file.

2009-07-08, 15:49	#38
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 3⁶×13 Posts	I had exit 11, when filtering (new style, immediately in memory) on a very small gnfs, on linux x86_64. The available memory was determined fairly correctly by get_ram_size(). Then some filtering and exit 11.

2009-07-08, 16:03	#39
jasonp Tribal Bullet Oct 2004 3,541 Posts	Sorry, I was referring to jrk's second bug report, with the problems during singleton removal. Shaopu, do you have the same problem now that the source has been updated?

2009-07-09, 18:27	#42
jasonp Tribal Bullet Oct 2004 3,541 Posts	jrk, I downloaded your relations but the crash doesn't happen with the latest code. Does it work for you now?

2009-07-10, 00:33	#44
jasonp Tribal Bullet Oct 2004 3,541 Posts	Grr, still no crash when run locally. Also no crash on my 64-bit system. What platform and compiler version are you using?