|
2009-11-17, 03:20
|
#23 |
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
947710 Posts |
The last crash is not very interesting.
All roots failes with these notes: Code:
Mon Nov 16 07:01:53 2009 reading relations for dependency 1 Mon Nov 16 07:01:55 2009 read 2745464 cycles Mon Nov 16 07:02:00 2009 cycles contain 8607833 unique relations Mon Nov 16 07:03:45 2009 read 8607833 relations Mon Nov 16 07:03:45 2009 number of relations is not even Mon Nov 16 07:03:46 2009 reading relations for dependency 2 Mon Nov 16 07:03:48 2009 read 2747401 cycles Mon Nov 16 07:03:53 2009 cycles contain 8610828 unique relations Mon Nov 16 07:05:36 2009 read 8610828 relations Mon Nov 16 07:05:56 2009 algebraic side is not a square! Mon Nov 16 07:05:56 2009 reading relations for dependency 3 2. Also check the .dat file for strange lines (but do not delete them, or the .dep will become incompatible). awk -F: 'NF!=3 && !(NF==2 && $2=="")' your...dat grep -v "^[-0-9A-Fa-f:,]" your...dat (Of course the first line should pop out of these tests. It is "N ..........".) 3. Did you sieve with the latest sievers? I wonder if this is because I "fixed" the "mpqs failed" bug. Symptoms: there will be relations that have the same primes in A or R sections. Try on your...dat file: Code:
#!/usr/bin/perl -w
while(<>) {
print $_ if /:([0-9a-f]+),\1/;
}
Last fiddled with by Batalov on 2009-11-17 at 03:26 Reason: while I was writing... doubled Greg's |
|
|
|
2009-11-17, 12:13
|
#24 | |
Jun 2003
Ottawa, Canada
3·17·23 Posts |
I will try an older binary and see what happens.
Your grep line produced nothing. Your awk line produced this: Code:
6271457,6292155:8bd91d1,1a50db01,2F45,9F7D,28AB6 -1284119,1398665:a4e6b6d,a3d37e9,2363,C9AF,1489C3,1933,5B3,5,7,7,47:1-50885,41286:1862ac23,7d0ee79,20FB,43403,14578D,1475395,11,2:d13ff5,ad18453,AD,5,5,5,43EA0D 28125995,8631287:2a9ec29,863b825,8551,CB29,FFF1,3FD1F,3,11,1F,101,35F:DE83,236EF3,314731222,798705:372ec05,73b52b5,FBD6B,1B83A1,3B63FB,1613,295,5,25:6e2d4a7,5CBF,72EF,D567,883,11,2B,2,2,2,2,5E4AB3 2655041,3207944:2288af7,44cee77,249B,28CC9,2D0BBF,1807,94D,3,3,7,47, 2144397,1 47612059,8753686:7aec3b7,8acbe3f,20309,342865,3E55FB,7CC073,3,43,2:395bc0d1410755,59527:d5510e3,1046cbdf,12FF5,130D27,273ECB,A75,BF,1AF:196b083,19AFE9,1C99,8BF,5,5,5,7,84885D -6528060,3363431:20cc017,10430c03,11DA1,32 -13580215,5018876:1c749a5b,165b8415,268 -20853679,5295320:b62d6fb,1fdde25b,5D79,34781,4E8819,3,5,13,1F,1B1,259,2, 17273208,17023583:10dd9799,531c2f3,15745,19775,1B2DB,1C57DBD,1EBB -14630629,2439270:8c957b1,ce8216d,ED65,1E0911,3BFFAD,9A204D,5,17,2:a609a156631,180874:826dfe9,d3c2973,2C93,118C7,C8207,1659F1,821,3,2:1c3c1b07,94872ad,6791,3,3,3,6FE47F 41901279,2127844:1896d3e7,c08cf51,61403,35C1AF,12149E1,59,8B,38F, I didn't do the sieving it was RSALS so I'm not sure what version of the siever it was. Jeff. Quote:
|
|
|
|
|
|
2009-11-17, 13:27
|
#25 |
|
Tribal Bullet
Oct 2004
3,541 Posts |
How would one get relations glued together like that? Appending a windows output file to a unix output file with different EOL characters? Appending to siever output that was interrupted so no final CR in the file? Lots of factorizations have a few lines like this; it doesn't explain the line with the dash though. That machine was probably super overclocked :)
Last fiddled with by jasonp on 2009-11-17 at 13:28 |
|
|
|
|
2009-11-17, 13:57
|
#26 | |
|
Jun 2003
Ottawa, Canada
3×17×23 Posts |
Quote:
An update: all the windows binaries including the one that Jason produced on sourceforge (done with cygwin?) crash, but the Linux 64bit binary that I built with the 1.43 code did not crash and just completed showing prp58 and prp100 factors. So the Linux code seems to handle this situation fine. I ran the 64bit Windows binary built with debug code, this is what I get. Variables: Code:
+ obj 0x0000000000649a50 {input=0x000000000012fbf0 "4119920635610832117027022146854969787051502053758103615653887536974906120680391680621246965735169915172641486888857683983247749384769531624818097656906464449" factors=0x0000000000000000 flags=1027 ...} msieve_obj *
+ fb 0x000000000012d100 {rfb={...} afb={...} } factor_base_t *
+ num_cycles_out 0x0000000000000000 unsigned int *
+ cycle_list_out 0x0000000000000000 la_col_t * *
+ num_relations_out 0x000000000012e234 unsigned int *
+ rlist_out 0x000000000012e278 relation_t * *
compress 0 unsigned int
dependency 36 unsigned int
+ cycle_list 0x0000000000000000 {data=??? weight=??? cycle={...} } la_col_t *
+ rlist 0xcccccccccccccccc {a=??? b=??? rel_index=??? ...} relation_t *
num_cycles 0 unsigned int
num_relations 3435973836 unsigned int
Code:
> msieve.exe!nfs_read_cycles(msieve_obj * obj=0x0000000000649a50, factor_base_t * fb=0x000000000012d100, unsigned int * num_cycles_out=0x0000000000000000, la_col_t * * cycle_list_out=0x0000000000000000, unsigned int * num_relations_out=0x000000000012e234, relation_t * * rlist_out=0x000000000012e278, unsigned int compress=0, unsigned int dependency=36) Line 622 + 0x8 bytes C msieve.exe!nfs_find_factors(msieve_obj * obj=0x0000000000649a50, mp_t * n=0x000000000012fa00, factor_list_t * factor_list=0x000000000012f1b0) Line 356 C msieve.exe!factor_gnfs(msieve_obj * obj=0x0000000000649a50, mp_t * n=0x000000000012fa00, factor_list_t * factor_list=0x000000000012f1b0) Line 160 + 0x1d bytes C msieve.exe!msieve_run_core(msieve_obj * obj=0x0000000000649a50, mp_t * n=0x000000000012fa00, factor_list_t * factor_list=0x000000000012f1b0) Line 149 + 0x1d bytes C msieve.exe!msieve_run(msieve_obj * obj=0x0000000000649a50) Line 241 + 0x1d bytes C msieve.exe!factor_integer(char * buf=0x000000000012fbf0, unsigned int flags=1027, char * savefile_name=0x0000000000645d61, char * logfile_name=0x0000000000645d4e, char * nfs_fbfile_name=0x0000000000645d29, unsigned int * seed1=0x000000000012fd94, unsigned int * seed2=0x000000000012fdb4, unsigned int max_relations=0, unsigned __int64 nfs_lower=0, unsigned __int64 nfs_upper=0, cpu_type cpu=cpu_core, unsigned int cache_size1=32768, unsigned int cache_size2=6291456, unsigned int num_threads=1, unsigned int mem_mb=0) Line 183 C msieve.exe!main(int argc=13, char * * argv=0x0000000000645c80) Line 515 C msieve.exe!__tmainCRTStartup() Line 266 + 0x19 bytes C msieve.exe!mainCRTStartup() Line 182 C kernel32.dll!000000007794f56d() [Frames below may be incorrect and/or missing, no symbols loaded for kernel32.dll] ntdll.dll!0000000077a83281() I'm not sure if this makes a difference but I think with this number I did -nc1 on Windows, then -nc2 on unix, and then -nc3 on Windows (crash) but when I tried -nc3 on Unix it worked. Would that produce any incompatible files in the process? Jeff. Last fiddled with by Jeff Gilchrist on 2009-11-17 at 14:01 |
|
|
|
|
|
2009-11-17, 14:09
|
#27 |
|
Tribal Bullet
Oct 2004
3,541 Posts |
The sourceforge binary was built using minGW with gcc 4.x
As long as the endianness between machines is the same, files can be generated on one OS and moved to another without issue. The only ascii file is the relation file, and the relation-reading code explicitly supports DOS and unix EOL characters simultaneously. Does the linux binary also have the 'odd parity' errors on dependencies? If not, that's somewhat worrying and indicates as 32-bit issue that's more serious than not cleaning up on a null pointer. See if changing the if() block at gnfs/relation.c:620 to Code:
if (num_cycles == 0) {
free(cycle_list);
if (num_cycles_out != NULL)
*num_cycles_out = 0;
if (cycle_list_out != NULL)
*cycle_list_out = NULL;
if (num_relations_out != NULL)
*num_relations_out = 0;
if (rlist_out != NULL)
*rlist_out = NULL;
return;
}
Last fiddled with by jasonp on 2009-11-17 at 14:18 |
|
|
|
|
2009-11-17, 14:34
|
#28 | |
|
Jun 2003
Ottawa, Canada
3·17·23 Posts |
Quote:
Code:
Tue Nov 17 07:15:46 2009 commencing square root phase Tue Nov 17 07:15:46 2009 reading relations for dependency 1 Tue Nov 17 07:15:47 2009 read 2745464 cycles Tue Nov 17 07:15:57 2009 cycles contain 8607572 unique relations Tue Nov 17 07:17:17 2009 read 8607572 relations Tue Nov 17 07:18:24 2009 multiplying 8607572 relations Tue Nov 17 07:27:09 2009 multiply complete, coefficients have about 209.62 million bits Tue Nov 17 07:27:11 2009 initial square root is modulo 33327901 Tue Nov 17 07:44:47 2009 reading relations for dependency 2 Tue Nov 17 07:44:48 2009 read 2747401 cycles Tue Nov 17 07:44:58 2009 cycles contain 8610710 unique relations Tue Nov 17 07:46:18 2009 read 8610710 relations Tue Nov 17 07:47:25 2009 multiplying 8610710 relations Tue Nov 17 07:56:10 2009 multiply complete, coefficients have about 209.70 million bits Tue Nov 17 07:56:12 2009 initial square root is modulo 33526301 Tue Nov 17 08:13:49 2009 sqrtTime: 3483 Tue Nov 17 08:13:49 2009 prp58 factor: X... Tue Nov 17 08:13:49 2009 prp100 factor: Y... Tue Nov 17 08:13:49 2009 elapsed time 00:58:05 |
|
|
|
|
|
2009-11-17, 14:49
|
#29 |
|
Tribal Bullet
Oct 2004
1101110101012 Posts |
Well, the code actually proceeded to the square root so odd parity didn't happen. Grrr.
Last fiddled with by jasonp on 2009-11-17 at 14:50 |
|
|
|
|
2009-11-17, 17:00
|
#30 | |
|
Jun 2003
Ottawa, Canada
49516 Posts |
Quote:
Code:
Tue Nov 17 11:49:33 2009 reading relations for dependency 35 Tue Nov 17 11:49:35 2009 read 2745153 cycles Tue Nov 17 11:49:46 2009 cycles contain 8604690 unique relations Tue Nov 17 11:52:37 2009 read 8604690 relations Tue Nov 17 11:53:15 2009 algebraic side is not a square! Tue Nov 17 11:53:18 2009 reading relations for dependency 36 Tue Nov 17 11:53:20 2009 read 0 cycles Tue Nov 17 11:53:20 2009 reading relations for dependency 37 Tue Nov 17 11:53:22 2009 read 0 cycles Tue Nov 17 11:53:22 2009 reading relations for dependency 38 Tue Nov 17 11:53:24 2009 read 0 cycles Tue Nov 17 11:53:24 2009 reading relations for dependency 39 Tue Nov 17 11:53:27 2009 read 0 cycles Tue Nov 17 11:53:27 2009 reading relations for dependency 40 Tue Nov 17 11:53:29 2009 read 0 cycles Tue Nov 17 11:53:29 2009 reading relations for dependency 41 Tue Nov 17 11:53:31 2009 read 0 cycles Tue Nov 17 11:53:31 2009 reading relations for dependency 42 Tue Nov 17 11:53:33 2009 read 0 cycles Tue Nov 17 11:53:33 2009 reading relations for dependency 43 Tue Nov 17 11:53:36 2009 read 0 cycles Tue Nov 17 11:53:36 2009 reading relations for dependency 44 Tue Nov 17 11:53:38 2009 read 0 cycles Tue Nov 17 11:53:38 2009 reading relations for dependency 45 Tue Nov 17 11:53:40 2009 read 0 cycles Tue Nov 17 11:53:40 2009 reading relations for dependency 46 Tue Nov 17 11:53:42 2009 read 0 cycles Tue Nov 17 11:53:43 2009 reading relations for dependency 47 Tue Nov 17 11:53:45 2009 read 0 cycles Tue Nov 17 11:53:45 2009 reading relations for dependency 48 Tue Nov 17 11:53:47 2009 read 0 cycles Tue Nov 17 11:53:47 2009 reading relations for dependency 49 Tue Nov 17 11:53:50 2009 read 0 cycles Tue Nov 17 11:53:50 2009 reading relations for dependency 50 Tue Nov 17 11:53:52 2009 read 0 cycles Tue Nov 17 11:53:52 2009 reading relations for dependency 51 Tue Nov 17 11:53:54 2009 read 0 cycles Tue Nov 17 11:53:54 2009 reading relations for dependency 52 Tue Nov 17 11:53:57 2009 read 0 cycles Tue Nov 17 11:53:57 2009 reading relations for dependency 53 Tue Nov 17 11:53:59 2009 read 0 cycles Tue Nov 17 11:53:59 2009 reading relations for dependency 54 Tue Nov 17 11:54:01 2009 read 0 cycles Tue Nov 17 11:54:01 2009 reading relations for dependency 55 Tue Nov 17 11:54:03 2009 read 0 cycles Tue Nov 17 11:54:04 2009 reading relations for dependency 56 Tue Nov 17 11:54:06 2009 read 0 cycles Tue Nov 17 11:54:06 2009 reading relations for dependency 57 Tue Nov 17 11:54:08 2009 read 0 cycles Tue Nov 17 11:54:08 2009 reading relations for dependency 58 Tue Nov 17 11:54:10 2009 read 0 cycles Tue Nov 17 11:54:10 2009 reading relations for dependency 59 Tue Nov 17 11:54:13 2009 read 0 cycles Tue Nov 17 11:54:13 2009 reading relations for dependency 60 Tue Nov 17 11:54:15 2009 read 0 cycles Tue Nov 17 11:54:15 2009 reading relations for dependency 61 Tue Nov 17 11:54:17 2009 read 0 cycles Tue Nov 17 11:54:17 2009 reading relations for dependency 62 Tue Nov 17 11:54:19 2009 read 0 cycles Tue Nov 17 11:54:19 2009 reading relations for dependency 63 Tue Nov 17 11:54:22 2009 read 0 cycles Tue Nov 17 11:54:22 2009 reading relations for dependency 64 Tue Nov 17 11:54:24 2009 read 0 cycles Tue Nov 17 11:54:24 2009 sqrtTime: 7636 Tue Nov 17 11:54:24 2009 elapsed time 02:07:26 Last fiddled with by Jeff Gilchrist on 2009-11-17 at 17:01 |
|
|
|
|
|
2009-11-17, 17:33
|
#31 | |
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
36·13 Posts |
Quote:
|
|
|
|
|
|
2009-11-17, 19:31
|
#32 | |
Jul 2003
So Cal
2·34·13 Posts |
Quote:
|
|
|
|
|
|
2009-11-17, 20:25
|
#33 | |
|
Jun 2003
Ottawa, Canada
22258 Posts |
Quote:
Sorry about that, but at least we tracked down a condition that msieve crashes on with corrupted data files that is now fixed.  Jeff. |
|
|
|
|
 |
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 1.42 feedback | Andi47 | Msieve | 167 | 2009-10-18 19:37 |
| Msieve 1.41 Feedback | Batalov | Msieve | 130 | 2009-06-09 16:01 |
