What happens when factors are found by TF or ECM

[bur]

I recently ran TF via a colab session and it found a factor. I didn't observe it when it happened, only afterwards from the results. From that same results page of my Primenet account I saw that it seems like the next TF candidate started immediately after the factor was found. Usually the TF takes 45 mins per candidate, here it was only 20 mins.

Does TF stop after a factor was found? If so, I guess there's a reason, but I can't think of any, k could increase by 1 and two factors could thus be within the same bit-level. Or not?

But maybe I just misread the results.


Similar question regarding ECM. If a factor is found, but the assignment consists of more than 1 curve, will the factor be reported immediately? For smaller numbers it makes sense to restart ECM with the smaller new co-factor, for the large Mersennes maybe not such much. So what does actually happen when a factor is found during an assignment?

[ATH]

If you are using mfaktc you can choose yourself in the mfaktc.ini settings:

# possible values for StopAfterFactor:
# 0: Do not stop the current assignment after a factor was found.
# 1: When a factor was found for the current assignment stop after the
#    current bitlevel. This makes only sense when Stages is enabled.
# 2: When a factor was found for the current assignment stop after the
#    current class.
#
# Default: StopAfterFactor=1

StopAfterFactor=1

It depends if you are here for the main GIMPS project of finding Mersenne Primes. Then you should stop after finding a factor, because then that candidate is done and not prime, and hurry on to the next candidate.

But if you are interested in the GIMPS "side" project of finding lots and lots of factors of Mersenne numbers, you should continue.

[lycorn]

Originally Posted by bur

Similar question regarding ECM. If a factor is found, but the assignment consists of more than 1 curve, will the factor be reported immediately? For smaller numbers it makes sense to restart ECM with the smaller new co-factor, for the large Mersennes maybe not such much. So what does actually happen when a factor is found during an assignment?

This excerpt from undoc.txt should answer it:

By default, ECM will stop when a new factor is found. You can have ECM stop or
continue searching for factors by using a value of zero or one in prime.txt:
ContinueECM=n

[bur]

edit: I realized it's obviously visible from the stats if someone completed that range after a factor was found.


Ok, I can see how that makes sense for the goal of finding primes. But isn't it a problem for the stats? At least in case of TF. From mersenne.ca it will seem as if that range was checked while actually it wasn't fully done. It's noticable when looking at the factors, but not whether or not someone else already redid that range to complete it.



On the other hand, ECM will in all likelyhood find that any potentially missed factor sooner or later anyway.