How long is too long?
 

C110, older Pentium M processor (1.6GHz), no CUDA, running polynomial selection for about 24 hours so far.

Is something wrong? And how would I know?

[QUOTE=ThomRuley;360311]C110, older Pentium M processor (1.6GHz), no CUDA, running polynomial selection for about 24 hours so far.

Is something wrong? And how would I know?[/QUOTE]
Nothing wrong. You don't have a timeout :smile: You can try to find a better and better poly, for years, if you like... (well... is not so efficient...)

For a comparison: Single core time for a C110 in the factor.log file of yafu (which uses msieve for poly selection) is set to about 3400 seconds (I used 1 core of a Core2Duo CPU to check just now). So, most probably, you have already a very good poly.:smile:

Advice: if msieve looks complicate for you, try using yafu for a while. It does the most of the things for you, it is much easier for a start.

The development version of Msieve has a fix for a bug that messes up the calculation of elapsed time in Windows; v1.51 unfortunately does not have the fix.

If you are still using factmsieve.py, you can set:

...
USE_CUDA = False
GPU_NUM = 0
[B]MSIEVE_POLY_TIME_LIMIT = 0[/B]
...

to a hard time in minutes. (0 means to use msieve's timing.) But, I would first look in the .log file to see how long msieve wants to search. It will say something like:

poly select deadline: 15264
time limit set to 4.24 CPU-hours

The 15264 is CPU-seconds