Double Check P-1

[PhilF]

I was just assigned a Double Check in which P-1 factoring had not been done. This is a new one on me. I guess this means if an exponent gets assigned to a machine that does not have enough memory to do the normal P-1 factoring then the machine assigned the double check does the P-1. Please correct me if I'm wrong.

I'm curious. What if I find a factor? I guess that means the exponent will go to the cleared list, and the first LL test will remain in the unverified list forever?

[marc]

It just gets moved from the 1LL Test to the Factored list and saves you having to do the double check. In http://mersenneforum.org/forumdisplay.php?f=30 people are trying to do the P-1 tests before they ever get to double-checking.

[dave_0273]

I am pretty sure that the only way that the p-1 test will not get done with the first LL test is if it is manually disabled (as described in undoc.txt). Even if the computer is REALLY slow, it will just run stage 1 (which only takes like 10mb of RAM). Then when it is assigned for the double check, even though only stage 1 has been done, it won't bother doing the p-1 again.

If you find a factor while doing the p-1, the exponent will be moved to factors.cmp and the corresponding lines in hrf3.txt will be deleated.

I think (but you better check with someone else just to be sure), if a factor is found, and primenet credit related to LL testing the exponent is lost.

Out of interest, what is the exponent. I want to know why mersenne-aries didn't do it for you. Unless it is a exponent manually released by George for a quick double check due to a scary error code......

[PhilF]

It is 14566549.

The only way I can think of to find out that this exponent needed P-1 done is to look for exponents in the unverified exponents list that are not in the P-1 factoring list. That sounds like it might be more trouble than it is worth for such a small exponent. Is there another way?

[dave_0273]

Ok... this is a bit of a mystery....

The exponent....14566549....

From status.txt

14566549 D   65               0.0   6.0  66.0  02-Jul-05 02:42  02-Jul-05 02:40

From pminus1.txt

14566549,160000,2040000

And just for the record, phil hasn't returned the p-1 result that is in pminus1.txt. The way that I know this is because the exponent was assigned to him on the 2nd July, where as pminus1.txt was out on the 1st July. Therefore the server knew that the exponent had been p-1ed before it was assigned on the 2nd (to phil) for a double-check.

Maybe a hic-cup in the server??

(Not any consolation to you, but at least I know now that mersenne-aries didn't miss an exponent. Also, just for the record, mersenne-aries never p-1ed this exponent, so I am assuming that it must have been done with the original p-1 test.)

[garo]

In rare cases in the past the Primenet server has messed up the P-1 status of a number. When I find a doublecheck that has not had the P-1 step done according to the server I usually make it a point to cross check in the PMINUS1.TXT

[cheesehead]

Originally Posted by dave_0273

I am pretty sure that the only way that the p-1 test will not get done with the first LL test is if it is manually disabled (as described in undoc.txt).

... or if someone's still using a Prime95 earlier than version 20, which introduced the automatic P-1 prior to L-L.