Duplicated problem here (across two messages), in case it helps:
[code]
math10@math10:~/Math/yafuTest$ ./yafu

11/11/12 00:17:28 v1.33 @ math10, System/Build Info:
Using GMP-ECM 6.4.3, Powered by GMP 5.0.4
detected Pentium(R) Dual-Core CPU E5200 @ 2.50GHz
detected L1 = 32768 bytes, L2 = 2097152 bytes, CL = 64 bytes
measured cpu frequency ~= 2499.942000
using 20 random witnesses for Rabin-Miller PRP checks

===============================================================
======= Welcome to YAFU (Yet Another Factoring Utility) =======
======= bbuhrow@gmail.com =======
======= Type help at any time, or quit to quit =======
===============================================================
cached 78498 primes. pmax = 999983

>> nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789)

Illegal instruction (core dumped)
[/code][code]
math10@math10:~/Math/yafuTest$ ./yafu 'nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789)'


Illegal instruction (core dumped)
[/code][code]
math10@math10:~/Math/yafuTest$ gdb ./yafu 'nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789)'
GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2) 7.4-2012.04
Copyright (C) 2012 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".
For bug reporting instructions, please see:
<http://bugs.launchpad.net/gdb-linaro/>...
Reading symbols from /home/math10/Math/yafuTest/yafu...done.
/home/math10/Math/yafuTest/nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789): No such file or directory.
(gdb) run
Starting program: /home/math10/Math/yafuTest/yafu


11/11/12 00:21:46 v1.33 @ math10, System/Build Info:
Using GMP-ECM 6.4.3, Powered by GMP 5.0.4
detected Pentium(R) Dual-Core CPU E5200 @ 2.50GHz
detected L1 = 32768 bytes, L2 = 2097152 bytes, CL = 64 bytes
measured cpu frequency ~= 2499.931500
using 20 random witnesses for Rabin-Miller PRP checks

===============================================================
======= Welcome to YAFU (Yet Another Factoring Utility) =======
======= bbuhrow@gmail.com =======
======= Type help at any time, or quit to quit =======
===============================================================
cached 78498 primes. pmax = 999983


>> nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789)


Program received signal SIGILL, Illegal instruction.
0x000000000045b904 in msieve_obj_new ()
(gdb) bt full
#0 0x000000000045b904 in msieve_obj_new ()
No symbol table info available.
#1 0x0000000000447f3d in nfs (fobj=0x881ee0) at factor/nfs/nfs.c:157
input = 0x893030 "1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789"
obj = <optimized out>
cpu = cpu_core
mpN = {nwords = 54001, val = {54011, 54013, 54037, 54049, 54059,
54083, 54091, 54101, 54121, 54133, 54139, 54151, 54163, 54167,
54181, 54193, 54217, 54251, 54269, 54277, 54287, 54293, 54311,
54319, 54323, 54331, 54347, 54361, 54367, 54371, 54377, 54401}}
factor_list = {num_factors = 48473, final_factors = {0xbd670000bd61,
0xbd710000bd6b, 0xbd8f0000bd8b, 0xbd9b0000bd95, 0xbdb30000bd9d,
0xbdcd0000bdbb, 0xbde30000bdd1, 0xbdef0000bdeb, 0xbe090000be07,
0xbe210000be15, 0xbe270000be25, 0xbe5d0000be5b, 0xbe750000be6f,
0xbe7f0000be79, 0xbe8d0000be8b, 0xbe9f0000be93, 0xbeb10000bea9,
0xbeb70000beb5, 0xbed90000becf, 0xbee50000bedb, 0xbef30000bee7,
0xbf0b0000bef9, 0xbf390000bf33, 0xbf5d0000bf4d, 0xbf6b0000bf5f,
0xbf7b0000bf71, 0xbf890000bf87, 0xbf930000bf8d, 0xbfad0000bfa1,
0xbfcf0000bfb9, 0xbfdd0000bfd5, 0xbfe30000bfe1, 0xc0050000bff3,
0xc0130000c011, 0xc0290000c019, 0xc0310000c02f, 0xc03b0000c037,
0xc0650000c047, 0xc07d0000c06d, 0xc0910000c07f, 0xc0b30000c09b,
0xc0bb0000c0b5, 0xc0d70000c0d3, 0xc0ef0000c0d9, 0xc1010000c0f1,
---Type <return> to continue, or q <return> to quit---
0xc1090000c103, 0xc1190000c115, 0xc1330000c12b, 0xc1450000c137,
0xc15b0000c149, 0xc1790000c173, 0xc1810000c17b, 0xc18d0000c18b,
0xc1bd0000c197, 0xc1cd0000c1c3, 0xc1e10000c1db, 0xc1ff0000c1e7,
0xc2050000c203, 0xc2210000c211, 0xc23f0000c22f, 0xc24d0000c24b,
0xc25d0000c253, 0xc27b0000c277, 0xc2890000c27d, 0xc2930000c28f,
0xc2a70000c29f, 0xc2bd0000c2b3, 0xc2d50000c2cf, 0xc2ff0000c2e3,
0xc3070000c301, 0xc3130000c311, 0xc3250000c317, 0xc3490000c347,
0xc3650000c34f, 0xc3710000c367, 0xc3830000c37f, 0xc3950000c385,
0xc3a70000c39d, 0xc3b50000c3ad, 0xc3c70000c3bf, 0xc3d10000c3cb,
0xc3e30000c3d3, 0xc3ef0000c3e9, 0xc41f0000c401, 0xc4330000c42d,
0xc4550000c437, 0xc4610000c457, 0xc4730000c46f, 0xc4910000c487,
0xc49d0000c499, 0xc4b70000c4a5, 0xc4c90000c4bb, 0xc4d30000c4cf,
0xc4f10000c4eb, 0xc5090000c4f7, 0xc51d0000c51b, 0xc5470000c541,
0xc55f0000c551, 0xc56f0000c56b, 0xc5770000c575, 0xc59b0000c595,
0xc5a10000c59f, 0xc5c30000c5a7, 0xc5db0000c5d7, 0xc5fb0000c5ef,
0xc6230000c613, 0xc6410000c635, 0xc6550000c64f, 0xc6650000c659,
0xc6910000c685, 0xc6a10000c697, 0xc6b30000c6a9, 0xc6cb0000c6b9,
0xc6dd0000c6cd, 0xc6f10000c6eb, 0xc70d0000c707, 0xc71b0000c719,
0xc7310000c72d, 0xc7570000c739, 0xc7670000c763, 0xc7750000c773,
0xc7a50000c77f, 0xc7bd0000c7bb, 0xc7cf0000c7c1, 0xc7e10000c7d5,
0xc7fd0000c7f9, 0xc8030000c7ff, 0xc81d0000c811, 0xc8290000c827,
0xc83f0000c839, 0xc8570000c853, 0xc8810000c86b, 0xc88f0000c88d,
0xc8950000c893, 0xc8b70000c8a1, 0xc8d50000c8cf, 0xc8dd0000c8db,
---Type <return> to continue, or q <return> to quit---
0xc8e70000c8e3, 0xc8ef0000c8ed, 0xc9050000c8f9, 0xc9170000c911,
0xc91f0000c919, 0xc9370000c92f, 0xc9410000c93d, 0xc95f0000c953,
0xc9790000c96b, 0xc9890000c97d, 0xc9970000c98f, 0xc9af0000c99d,
0xc9bf0000c9b5, 0xc9d90000c9cb, 0xc9e30000c9df, 0xca010000c9eb,
0xca090000ca07, 0xca370000ca25, 0xca4b0000ca39, 0xca5b0000ca55,
0xca730000ca69, 0xca7f0000ca75, 0xca930000ca8d, 0xca9f0000ca9d,
0xcabb0000cab5, 0xcac90000cac3, 0xcae50000cad9, 0xcb030000caed,
0xcb090000cb05, 0xcb290000cb17, 0xcb3b0000cb35, 0xcb590000cb53,
0xcb650000cb63, 0xcb870000cb71, 0xcb9f0000cb99, 0xcbb90000cbb3,
0xcbd10000cbc3, 0xcbd70000cbd5, 0xcbe90000cbdd, 0xcc0d0000cbff,
0xcc1d0000cc19, 0xcc2b0000cc23, 0xcc430000cc41, 0xcc590000cc4d,
0xcc890000cc61, 0xcc910000cc8b, 0xcca30000cc9b, 0xccd10000cca7,
0xcce90000cce5, 0xcd150000cd09, 0xcd250000cd1f, 0xcd3d0000cd31,
0xcd490000cd3f, 0xcd570000cd51, 0xcd630000cd5b, 0xcd810000cd67,
0xcd970000cd93, 0xcdbb0000cd9f, 0xcdd30000cdc1, 0xcde50000cdd9,
0xcdf10000cde7, 0xcdfd0000cdf7, 0xce150000ce0b...}}
flags = <optimized out>
job = {fblim = 0, lpb = 0, mfb = 0, lambda = 0, siever = 54727,
qrange = 54751,
sievername = "\357\325\000\000\365\325\000\000\373\325\000\000\003\326\000\000\017\326\000\000-\326\000\000\061\326\000\000C\326\000\000U\326\000\000]\326\000\000a\326\000\000{\326\000\000\205\326\000\000\207\326\000\000\235\326\000\000\245\326\000\000\257\326\000\000\275\326\000\000\001\000\000\000\000\000\0---Type <return> to continue, or q <return> to quit---
[/code]At this point the output contained characters that wouldn't copy/paste. The rest of the output is in the following post

Continuation after uncopyable section:
[code]
min_rels = 32767, current_rels = 0, type = 0, size = 0,
poly_time = 0, last_leading_coeff = 4}
relations_needed = <optimized out>
last_specialq = 0
stop = {tv_sec = 234251811345675, tv_usec = 234329120757011}
start = {tv_sec = 1352611401, tv_usec = 817371}
bstop = {tv_sec = 233942573700273, tv_usec = 234062832784605}
bstart = {tv_sec = 233736415270035, tv_usec = 233830904550565}
difference = <optimized out>
t_time = <optimized out>
pre_batch_rels = <optimized out>
tmpstr = '\000' <repeats 28 times>, "\026\000\000\000\000\354\000\000\000\000\000\000\000\000\000\000\270\336\377\377\000\000\000\000\271\354", '\000' <---Type <return> to continue, or q <return> to quit---
repeats 11 times>"\354, \000\000\317\354\000\000\377\377\377\377\377\377\377\377\324\325\377\377\001\000\000\000\000\000\000\000\262\336\377\377\211\020\034\213\000\000\000\000\000\000\000\000\377\377\377\377\377\377\377\377\002\000\000\000\v\225]\000\000\000\000\000\377\377\377\377\000\000\000\000\n", '\000' <repeats 19 times>"\377, \177\000\000\000\000\000\000\000\000\000\000zw\210\000\000\000\000\000]\000\000\000\000\000\000\000\350\225]", '\000' <repeats 13 times>"\220, \366\207\000\000\000\000\000u\000\000\000\000\000\000\000\347\225]", '\000' <repeats 13 times>"\270, \336\377\377\026\000\000\000\240-\210\000\000\000\000\000\245\224]", '\000' <repeats 13 times>, "hB\210\000\026", '\000' <repeats 19 times>, "M\357\000\000\020\000\000\000k\357\000\000q\357\000\000\000\000\000\000\225"...
process_done = <optimized out>
nfs_state = <optimized out>
#2 0x00000000004133cd in feval (func=58, nargs=<optimized out>, fobj=0x881ee0)
at top/calc.c:2309
mp1 = {val = 0x891c30, alloc = 100, size = 1, type = 3}
mp2 = {val = 0x891f60, alloc = 100, size = 1, type = 3}
mp3 = {val = 0x892290, alloc = 100, size = 1, type = 3}
tmp1 = {val = 0x8925c0, alloc = 100, size = 1, type = 3}
tmp2 = {val = 0x8928f0, alloc = 100, size = 1, type = 3}
gmpz = {{_mp_alloc = 19, _mp_size = 0, _mp_d = 0x4}}
str = {s = 0x892c20 "", nchars = 1, alloc = 1024}
i = 0
---Type <return> to continue, or q <return> to quit---
n64 = <optimized out>
j = <optimized out>
k = <optimized out>
t = <optimized out>
tstart = {tv_sec = 8942704, tv_usec = 4496518}
tstop = {tv_sec = 8979712, tv_usec = 8895328}
difference = <optimized out>
lower = <optimized out>
upper = <optimized out>
inc = <optimized out>
count = <optimized out>
#3 0x00000000004145c9 in calc (in=0x7fffffffe150, fobj=0x881ee0)
at top/calc.c:963
i = <optimized out>
retval = 0
na = 1
func = 58
j = <optimized out>
k = <optimized out>
stk = {elements = 0x88ac40, num = 0, size = 20, top = -1, type = 0}
tmp = 0x8900b0
post = 0x891800
tokens = 0x88a910
---Type <return> to continue, or q <return> to quit---
tok = <optimized out>
strN = <optimized out>
delim = " "
token_types = 0x887290
num_tokens = 4
varstate = <optimized out>
tmpz = {val = 0x8914d0, alloc = 100, size = 5, type = 3}
#4 0x000000000040a609 in main (argc=<optimized out>, argv=<optimized out>)
at top/driver.c:309
insize = 1024
input_exp = 0x87f690 "ans"
ptr = <optimized out>
indup = 0x87faa0 ""
str = {
s = 0x87feb0 "nfs(1234567891234567891235467891234567891234567891234567891234567912345678912345679123456789)", nchars = 94, alloc = 1024}
tmp = {val = 0x887750, alloc = 100, size = 1, type = 3}
nooutput = 0
offset = <optimized out>
slog = 1
is_cmdline_run = <optimized out>
logfile = 0x887510
fobj = 0x881ee0
(gdb)
[/code]

[QUOTE=Dubslow;317870]That's not good :razz:

I just tested my copy of that binary, and AFAICT (tested GNFS poly select and GNFS sieving) it works fine.

Here's a test: if your command that's crashing is `./yafu blah blah blah`, please try running `gdb ./yafu blah blah blah`, and then when that program starts, type `run`; when YAFU crashes, type `bt full` and post the output here.

Unless B[sup]2[/sup] has better insight?

Perhaps gcc only compiled it for my architecture? But then the ecm compile of mine you're using wouldn't work either...[/QUOTE]
If it's still needed after EdH's input, I can probably try gdb later today.

Just some additional info:

I tried it on both the Pentium (shown) and an AMD. Both were dual core and both failed with Dubslow's binary, but worked from the svn, which was compiled against revision 722 of msieve.

Also, in case it's of interest, here's a screen shot of the section that wouldn't copy/paste:

Was the binary in question built with -march=native? Could there be an instruction set incompatibility with the machines where it crashes? That could be an issue. Prefetches are different between AMD and Intel, as are some of the vector instructions.

[QUOTE=jasonp;317925]Was the binary in question built with -march=native? Could there be an instruction set incompatibility with the machines where it crashes? That could be an issue. Prefetches are different between AMD and Intel, as are some of the vector instructions.[/QUOTE]

I made sure to disable it.

:doh!: I made sure to disable it in YAFU, but not the Msieve compile.... :doh!: :bangheadonwall: Sorry guys.

[QUOTE=Dubslow;317934]I made sure to disable it.

:doh!: I made sure to disable it in YAFU, but not the Msieve compile.... :doh!: :bangheadonwall: Sorry guys.[/QUOTE]
Works great now. Thanks!

Factor 1024-bit on 2x512-bit
 

need to find a pair of 512-bit numbers that when multiplied will give 1024 bit number, similar to a 1024 bit number given in the first 750 binary digits
(is there some function in yafu? or some one know algoritm for this task)

[QUOTE=Ogonia;318031]need to find a pair of 512-bit numbers that when multiplied will give 1024 bit number, similar to a 1024 bit number given in the first 750 binary digits
(is there some function in yafu? or some one know algoritm for this task)[/QUOTE]

Responses moved [URL="http://www.mersenneforum.org/showthread.php?t=17424"]here [/URL]to keep everything in one place...

Stupid question:

My computer crashed last night for reasons not related to yafu. You know how the mother life is a bitch, so in that moment (according with the nfs.log) yafu was doing linear algebra on [URL="http://factordb.com/sequences.php?se=1&aq=731424&action=range&fr=803&to=803"]this C131[/URL]. Last lines in the file were:

[CODE]linear algebra at 0.1%, ETA 0h56m
checkpointing every 1460000 dimensions
[/CODE]

No idea how "advanced" into it it was.
After restart, it went to do "filtering" and it did filtering again for the ~~30M relations, which took about 5 minutes, and then, the last two lines in the files said:

[CODE]linear algebra at 0.1%, ETA 0h[COLOR=Red]51[/COLOR]m [COLOR=Red](??)[/COLOR]
checkpointing every [COLOR=Red]1220000[/COLOR] dimensions
[/CODE]

Is it not the same set of relations? Did the first "filtering" modified them somehow? (the file time is still from yesterday, it doesnt seem to be modified). Is the "filtering" process a kind of "randomized"? I was thinking this is very deterministic, but it may be that my understanding of NFS is still at low level.

And the strangest of all, on the screen, since I am writing this, I have:
[CODE]linear algebra completed xxxxxx of [COLOR=Red]1191396[/COLOR] dimensions (xx.x%, ETA 0h30m)[/CODE]
(i.e. under 1M22 assumed to be "checkpointed"). Why so? [COLOR=White](grr, this Raman is contagious!)
[/COLOR]

The LA does a little work to start with, measures how long it takes, then calibrates the checkpoint interval so that checkpoints happen about once per hour irrespective of how fast the machine is (at least that's how it's supposed to work). But the calibration phase is affected by how loaded the machine is, so I would expect some variation there. The calibration is repeated when you restart from checkpoint too.

The filtering is supposed to be deterministic, but I can't guarantee it produces the same matrix twice in a row because the work it does is so damn complex. The LA always starts from a randomly generated vector, so even with an identical starting matrix it's producing different numbers every time.