CADO NFS - Page 7

frmky · 2009-06-13, 06:16

After the successful small GNFS run, I tried a larger SNFS run with the same binaries, but no luck:

Code:

Multiply ker and character matrix
64 rows done      
Computing tiny kernel
dim of ker = 63
Sorry 0-th vector is 0
Sorry 1-th vector is 0
Sorry 2-th vector is 0
Sorry 3-th vector is 0
Sorry 4-th vector is 0
Sorry 5-th vector is 0
Sorry 6-th vector is 0
Sorry 7-th vector is 0
Sorry 8-th vector is 0
Sorry 9-th vector is 0
Sorry 10-th vector is 0
Sorry 11-th vector is 0
Sorry 12-th vector is 0
Sorry 13-th vector is 0
Sorry 14-th vector is 0
Sorry 15-th vector is 0
Sorry 16-th vector is 0
Sorry 17-th vector is 0
Sorry 18-th vector is 0
Sorry 19-th vector is 0
Sorry 20-th vector is 0
Sorry 21-th vector is 0
Sorry 22-th vector is 0
Sorry 23-th vector is 0
Sorry 24-th vector is 0
Sorry 25-th vector is 0
Sorry 26-th vector is 0
Sorry 27-th vector is 0
Sorry 28-th vector is 0
Sorry 29-th vector is 0
Sorry 30-th vector is 0
Sorry 31-th vector is 0
Sorry 32-th vector is 0
Sorry 33-th vector is 0
Sorry 34-th vector is 0
Sorry 35-th vector is 0
Sorry 36-th vector is 0
Sorry 37-th vector is 0
Sorry 38-th vector is 0
Sorry 39-th vector is 0
Sorry 40-th vector is 0
Sorry 41-th vector is 0
Sorry 42-th vector is 0
Sorry 43-th vector is 0
Sorry 44-th vector is 0
Sorry 45-th vector is 0
Sorry 46-th vector is 0
Sorry 47-th vector is 0
Sorry 48-th vector is 0
Sorry 49-th vector is 0
Sorry 50-th vector is 0
Sorry 51-th vector is 0
Sorry 52-th vector is 0
Sorry 53-th vector is 0
Sorry 54-th vector is 0
Sorry 55-th vector is 0
Sorry 56-th vector is 0
Sorry 57-th vector is 0
Sorry 58-th vector is 0
Sorry 59-th vector is 0
Sorry 60-th vector is 0
Sorry 61-th vector is 0
Sorry 62-th vector is 0

alex_148 · 2009-07-11, 22:33

Hello! I have some errors while compile under cygwin: it's like that

cantor/mpfq_2_128.h.32:791:error:invalid operands of types 'long long int _vector_' and 'long long int _vector_' to binary 'operator^'.

jasonp · 2009-07-27, 19:15

Congratulations to the CADO group for making their repository available!

henryzz · 2010-07-17, 11:15

Code:

:~/Desktop/cado-nfs/cado-nfs/trunk$ ./run_example.sh
Testing factorization as given by ./params/params.c59 in /tmp/cado.xaXZdQgYzS
./cadofactor.pl cadodir=/home/david/Desktop/cado-nfs/cado-nfs/trunk/build/Jimmy-Ubuntu /tmp/cado.xaXZdQgYzS/param machines=/tmp/cado.xaXZdQgYzS/mach_desc wdir=/tmp/cado.xaXZdQgYzS delay=60 sievenice=0 selectnice=0 logfile=/tmp/cado.xaXZdQgYzS/out
Info:--------------------------------------------------------------------------
Info:Initialization
Info:--------------------------------------------------------------------------
Info:Reading the parameters...
Info:Reading the machine description file...
Info:Initializing the working directory...
Info:--------------------------------------------------------------------------
Info:Polynomial selection
Info:--------------------------------------------------------------------------
Info:No job status file found. Creating empty one.
Info:Starting new jobs...
    Info:Sending `c59.n' to `localhost'...
    Info:Starting job: c59 localhost        1        100001  
    Info:Starting job: c59 localhost        100001   2e5     
Info:Total interval coverage:   0 %.
Info:Waiting for 60 seconds before checking again...
Info:Checking all running jobs...
    Info:Checking job: c59 localhost        1        100001  
        Info:Running...
    Info:Checking job: c59 localhost        100001   2e5     
        Info:Running...
Info:Starting new jobs...
Info:Total interval coverage:   0 %.
Info:Waiting for 60 seconds before checking again...
Info:Checking all running jobs...
    Info:Checking job: c59 localhost        1        100001  
        Info:Finished!
    Info:Checking job: c59 localhost        100001   2e5     
        Info:Finished!
Info:Retrieving job data...
    Info:Retrieving `c59.kjout.1-100001' from `localhost'...
    Info:Retrieving `c59.kjout.100001-2e5' from `localhost'...
Info:Starting new jobs...
Info:Total interval coverage: 100 %.
Info:Cleaning up...
Info:All done!
Info:The best polynomial is from `c59.kjout.1-100001' (E = 17.32).
Info:Generating factor base...
Info:Computing free relations...
Info:--------------------------------------------------------------------------
Info:Sieve
Info:--------------------------------------------------------------------------
Info:Checking previous files...
    Info:Imported 750 relations from `c59.freerels.gz'.
Info:No job status file found. Creating empty one.
Info:Starting new jobs...
    Info:Sending `c59.poly' to `localhost'...
    Info:Sending `c59.roots' to `localhost'...
    Info:Starting job: c59 localhost        400000   405000  
Info:Running total: 750 relations.
Info:Waiting for 60 seconds before checking again...
Info:Checking all running jobs...
    Info:Checking job: c59 localhost        400000   405000  
        Info:Running...
Info:Starting new jobs...
Info:Running total: 750 relations.
Info:Waiting for 60 seconds before checking again...
Info:Checking all running jobs...
    Info:Checking job: c59 localhost        400000   405000  
        Info:Finished!
Info:Retrieving job data...
    Info:Retrieving `c59.rels.400000-405000.gz' from `localhost'...
        Info:Imported 287703 relations from `c59.rels.400000-405000.gz'.
Info:Starting new jobs...
    Info:Starting job: c59 localhost        405000   410000  
Info:Running total: 288453 relations.
Info:--------------------------------------------------------------------------
Info:Duplicate and singleton removal
Info:--------------------------------------------------------------------------
Info:Removing duplicates...
    Info:split new files in 4 slices...
    Info:removing duplicates on slice 0...
    Info:removing duplicates on slice 1...
    Info:removing duplicates on slice 2...
    Info:removing duplicates on slice 3...
    Info:Number of relations left: 288207.
Info:Removing singletons...
    Info:Not enough relations! Continuing sieving...
Info:Waiting for 60 seconds before checking again...
Info:Checking all running jobs...
    Info:Checking job: c59 localhost        405000   410000  
        Info:Finished!
Info:Retrieving job data...
    Info:Retrieving `c59.rels.405000-410000.gz' from `localhost'...
        Info:Imported 283213 relations from `c59.rels.405000-410000.gz'.
Info:Starting new jobs...
    Info:Starting job: c59 localhost        410000   415000  
Info:Running total: 571666 relations.
Info:--------------------------------------------------------------------------
Info:Duplicate and singleton removal
Info:--------------------------------------------------------------------------
Info:Removing duplicates...
    Info:split new files in 4 slices...
    Info:removing duplicates on slice 0...
    Info:removing duplicates on slice 1...
    Info:removing duplicates on slice 2...
    Info:removing duplicates on slice 3...
    Info:Number of relations left: 570679.
Info:Removing singletons...
    Info:Nrows: 28740; Ncols: 28580; Excess: 160.
Info:Join all no duplicates files into one file...
    Info:clean directory nodup...
Info:Cleaning up...
    Info:Killing job:  c59 localhost        410000   415000  
        Info:Truncating `c59.rels.410000-415000' to range 410000-412860...
        Info:Imported 176451 relations from `c59.rels.410000-412860.gz'.
Info:All done!
Info:--------------------------------------------------------------------------
Info:Merge
Info:--------------------------------------------------------------------------
Info:Merging relations...
    Info:Minimal bwcost: 6262176018.
Info:Replaying merge history...
    Info:Nrows: 9875; Ncols: 9715; Weight: 969672.
Info:--------------------------------------------------------------------------
Info:Linear algebra
Info:--------------------------------------------------------------------------
Warning:Parameter 'skip' currently unhandled by bwc code
Info:Calling Block-Wiedemann (new code)...
    Error:Command `/home/david/Desktop/cado-nfs/cado-nfs/trunk/build/Jimmy-Ubuntu/linalg/bwc/bwc.pl :complete seed=1 thr=2x2 mpi=1x1 matrix=/tmp/cado.xaXZdQgYzS/c59.small nullspace=left mm_impl=bucket interleaving=0 interval=100 mode=u64 mn=64 splits=0,64 ys=0..64 wdir=/tmp/cado.xaXZdQgYzS/c59.bwc bwc_bindir=/home/david/Desktop/cado-nfs/cado-nfs/trunk/build/Jimmy-Ubuntu/linalg/bwc >> /tmp/cado.xaXZdQgYzS/c59.bwc.stderr 2>&1' terminated unexpectedly with exit status 1.
FAILED ; data left in /tmp/cado.xaXZdQgYzS

anyone know how to fix this error?

BWetter246 · 2012-06-03, 21:13

When i try to compile cado-nfs from the repository, i get an error at ./utils/cachesize_cpuid.c:47:3. Inconsistent operand constraints in an 'asm'. Does anyone know what this error means or how to fix this? I am using gcc 4.6.3 (Ubuntu/Linaro 4.6.3).

BWetter246 · 2012-06-03, 23:17

ok, so i pretty much changed the cpuid function in cachesize_cpuid.c and it compiled.

Code:

#include <cpuid.h>
void cpuid(uint32_t res[4], uint32_t op) {
  __get_cpuid(op, &res[0], &res[1], &res[2], &res[3]);
}

jasonp · 2012-06-04, 02:04

Looking at the original, it appears that eax is being referenced by name as an input and an output to the asm, which older versions of gcc inline asm syntax allowed but current versions do not. I'll forward the report to Paul's group; it's strange they haven't seen it already.

BWetter246 · 2012-06-04, 16:17

how about this

Code:

void cpuid(uint32_t res[4], uint32_t op) {
#ifdef	__GNUC__
  __asm__ volatile( "pushl %%ebx   \n\t"
                "cpuid         \n\t"
                "movl %%ebx, %1\n\t"
                "popl %%ebx    \n\t"
                : "=a"(res[0]), "=r"(res[1]), "=c"(res[2]), "=d"(res[3])
                : "a"(op)
                : "cc" );
#else
#error "Please teach your compiler how to call cpuid"
#endif
}

jasonp · 2012-06-05, 11:09

Paul writes:

Quote:

since all CADO-NFS developers are not on mersenneforum, in general it is better to advice people to report problems on the CADO-NFS bug tracker

Regarding your CPUID substitute, the problem I see is that you have a "=a" in the output list and an "a" in the input list. While the syntax does let you reuse a specific register, I didn't think mentioing it by name twice was allowed. Msieve does something similar:

Code:

	#define CPUID(code, a, b, c, d) 			\
		ASM_G volatile(					\
			"movl %%ebx, %%esi   \n\t"		\
			"cpuid               \n\t"		\
			"movl %%ebx, %1      \n\t"		\
			"movl %%esi, %%ebx   \n\t"		\
			:"=a"(a), "=m"(b), "=c"(c), "=d"(d) 	\
			:"0"(code) : "%esi")

bai · 2012-06-05, 14:10

Quote:

Originally Posted by BWetter246

When i try to compile cado-nfs from the repository, i get an error at ./utils/cachesize_cpuid.c:47:3. Inconsistent operand constraints in an 'asm'. Does anyone know what this error means or how to fix this? I am using gcc 4.6.3 (Ubuntu/Linaro 4.6.3).

BWetter&jason, thanks for pointing the bug and fix. As jason's post, does it work if we change the input line from

Quote:

: "a" (op)

to

Quote:

: "0" (op)

Btw: I just tested it (without above change) on a Ubuntu 12.04 machine (Ubuntu/Linaro 4.6.3) and it compiles fine. What flag do you use?

BWetter246 · 2012-06-05, 20:43

I used make in bash and it generated -g -W -Wall -O2.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CADO-NFS on windows	jux	CADO-NFS	25	2021-07-13 23:53
CADO help	henryzz	CADO-NFS	4	2017-11-20 15:14
CADO and WinBlows	akruppa	Programming	22	2015-12-31 08:37
CADO-NFS	skan	Information & Answers	1	2013-10-22 07:00
CADO	R.D. Silverman	Factoring	4	2008-11-06 12:35

2009-07-11, 22:33	#68
alex_148 Jul 2009 1 Posts	Hello! I have some errors while compile under cygwin: it's like that cantor/mpfq_2_128.h.32:791:error:invalid operands of types 'long long int _vector_' and 'long long int _vector_' to binary 'operator^'.

2009-07-27, 19:15	#69
jasonp Tribal Bullet Oct 2004 3,541 Posts	Congratulations to the CADO group for making their repository available!

2012-06-03, 21:13	#71
BWetter246 Aug 2005 17 Posts	When i try to compile cado-nfs from the repository, i get an error at ./utils/cachesize_cpuid.c:47:3. Inconsistent operand constraints in an 'asm'. Does anyone know what this error means or how to fix this? I am using gcc 4.6.3 (Ubuntu/Linaro 4.6.3).

2012-06-03, 23:17	#72
BWetter246 Aug 2005 10001₂ Posts	ok, so i pretty much changed the cpuid function in cachesize_cpuid.c and it compiled. Code: #include <cpuid.h> void cpuid(uint32_t res[4], uint32_t op) { __get_cpuid(op, &res[0], &res[1], &res[2], &res[3]); }

2012-06-04, 02:04	#73
jasonp Tribal Bullet Oct 2004 3,541 Posts	Looking at the original, it appears that eax is being referenced by name as an input and an output to the asm, which older versions of gcc inline asm syntax allowed but current versions do not. I'll forward the report to Paul's group; it's strange they haven't seen it already.

2012-06-05, 20:43	#77
BWetter246 Aug 2005 17 Posts	I used make in bash and it generated -g -W -Wall -O2.