Error Messages from GGNFS

[EdH]

Running Fedora 11, ggnfs via aliqueit and I have had quite a bit of fun getting it this far, but now:

polynomial selection complete
R0: -42273132966527721391
R1:  16736524361
A0: -1466629204024657642212000
A1:  361869072258352416960
A2:  15921217191190922
A3: -1856015274098
A4: -34285293
A5:  840
skew 23527.83, size 1.277516e-09, alpha -4.790957, combined = 2.993255e-09
elapsed time 00:23:44
sh: -c: line 0: unexpected EOF while looking for matching `"'
sh: -c: line 1: syntax error: unexpected end of file
WARNING: gnfs failed to find a factor. This really shouldn't happen.
I'll just run ecm till the end of time or a factor turns up...
Let's hope you don't run out of disk space before either of those.

I'm thinking the WARNING is from Aliqueit, but I'm not sure if the EOF errors are from ggnfs or something it called. gmp-ecm and msieve seem to be running fine. yafu is not running on this machine, but I have aliqueit set up to prefer msieve and have remarked out the yafu command line.

This seems like the last little hitch to finally having it all run, but these may just be obscuring a new set.

Any thoughts would be appreciated...

[schickel]

Originally Posted by EdH

Running Fedora 11, ggnfs via aliqueit and I have had quite a bit of fun getting it this far, but now:

WARNING: gnfs failed to find a factor. This really shouldn't happen.
I'll just run ecm till the end of time or a factor turns up...
Let's hope you don't run out of disk space before either of those.

I'm thinking the WARNING is from Aliqueit

....

Any thoughts would be appreciated...

Yes, the warning and the "end of time" comment are from aliqueit; there were no factors produced from the GGNFS run, so aliqueit gives up and falls back to just running ECM until it gets lucky (or you stop it).

As far as the issue with the NFS portion, the poly selection finished, so the process is at least getting started. My initial guess would be that there is still a bit of configuration that has to be set. Have you tried running a GNFS job outside of aliqueit? If you can do that, that would indicate that you have the configuration set up all correctly for GNFS and the problem is with aliqueit......

[EdH]

Originally Posted by schickel

Yes, the warning and the "end of time" comment are from aliqueit; there were no factors produced from the GGNFS run, so aliqueit gives up and falls back to just running ECM until it gets lucky (or you stop it).

As far as the issue with the NFS portion, the poly selection finished, so the process is at least getting started. My initial guess would be that there is still a bit of configuration that has to be set. Have you tried running a GNFS job outside of aliqueit? If you can do that, that would indicate that you have the configuration set up all correctly for GNFS and the problem is with aliqueit......

Thanks,

I only had a few minutes to play today and found that I have problems across all my computers. For some reason, on my linux machines, my factMsieve.pl references gnfs-lasieve4I11e, which my copy straight from the ggnfs beginners page (the copy that includes the factMsieve.pl, when compiled) doesn't include. I seem to be having other troubles that are piling up as well. so I'm going to step back and work out one machine at a time. Disappointingly, I found my laptop (winXP) ggnfs did "an illegal operation" during the night, so it quit as well. A second try yielded the same result; imagine that. So, basically, I reset aliqueit to use msieve and skip ggnfs on all my machines until I can get each figured out.

I'll still run the 700k stuff I reserved up to 100 digits, but it will be without ggnfs for now.

I do appreciate the help.

Take Care,
Ed

[schickel]

Originally Posted by EdH

Thanks,

I only had a few minutes to play today and found that I have problems across all my computers.

OK, well we can tackle everything one issue at a time 'til we get you up and running....

For some reason, on my linux machines, my factMsieve.pl references gnfs-lasieve4I11e, which my copy straight from the ggnfs beginners page (the copy that includes the factMsieve.pl, when compiled) doesn't include.

If you're using the factmsieve.pl from Jeff's page, don't worry, *sieve11e is only used for the smaller jobs (<95 digits), so as long as you set the NFS cut-off in aliqueit higher than that, there's no problem.

I seem to be having other troubles that are piling up as well. so I'm going to step back and work out one machine at a time. Disappointingly, I found my laptop (winXP) ggnfs did "an illegal operation" during the night, so it quit as well. A second try yielded the same result; imagine that.

That may be an isloated problem. Over the years, I have had jobs crop up once in a while that will crash on a particular special-q range (reproducible) which I've solved by just skipping the range that was crashing (change the q0 value in the job file and restart).

I think this is an example of an extreme corner case that generates incorrect values during the sieve. Again, very isolated, very easy to skip the bad range, just losing some hours if it happens while you're at work....

So, basically, I reset aliqueit to use msieve and skip ggnfs on all my machines until I can get each figured out.

I'll still run the 700k stuff I reserved up to 100 digits, but it will be without ggnfs for now.

Actually, depending on your core speed and luck, 100 + a little probably works. I made it up to 110 with one sequence without encountering a single NFS job (msieve/NFS cut-over set at 105 while I got the system setup)

I do appreciate the help.

Take Care,
Ed

Hey, that's what we're here for (well, some most of us, anyway).

[EdH]

Thanks schickel,

I think I have the linux machines running ggnfs right now. I had commented out the 4I11e references before reading your reply. It sounds like it would be more efficient to set the aliqueit value higher, but to me it would seem odd to have a line that references a missing file in the script. Actually, in my ggnfs folder, within a sub-sub-folder, I found a copy of the 4I11e program. I might copy it to the other location and see what happens.

Speaking of aliqueit, that is where my EOF errors were coming from, instead of ggnfs. I had a " sitting at the end of the ggnfs command line.

Unfortunately, while I was rejoicing with how my linux machines were seeming to work properly now, my WinXP machine was "arresting" ggnfs for doing "illegal operations."

Now I need to run that down, but I'll probably wait until it finishes taking its current job to 100 digits (99 now). For that one I did adjust the cut-off to over 100.

Thanks again for the assistance.

Happy New Year,
Ed