How can one benchmark P-1 testing?

[masser]

If you want to benchmark P-1 testing on your system, then you should be able to reserve 50 P-1 assignments from the manual reservations page.

Don't worry about the calculator at mersenne.ca; mprime is fairly sophisticated and will choose intelligent bounds.

As your machine progresses through the 50 P-1 assignments, experiment with changing the allocated RAM. Don't change the settings too often though; maybe every 10 completed assignments or so.

If you share the results here (exponent, total P-1 time, memory allocated for test), we may have suggestions for you.

[drkirkby]

Originally Posted by masser

If you want to benchmark P-1 testing on your system, then you should be able to reserve 50 P-1 assignments from the manual reservations page.

Don't worry about the calculator at mersenne.ca; mprime is fairly sophisticated and will choose intelligent bounds.

As your machine progresses through the 50 P-1 assignments, experiment with changing the allocated RAM. Don't change the settings too often though; maybe every 10 completed assignments or so.

If you share the results here (exponent, total P-1 time, memory allocated for test), we may have suggestions for you.

50 P-1 tests will take an awful long time, with zero chance of finding a prime.

Is it possible to force a P-1 test on exponents one has a PRP test for? For example, I will start the category 0 exponent 102529169 in a couple of days.

https://www.mersenne.org/report_expo...2529169&full=1

It has a P-1 test, but with B1=1 000 000, B2=30 000 000. Is it possible to force another P-1 test, to higher values of B2, by exploiting my memory? If it is, the obvious next question to ask is whether it is sensible to do so or not.

Dave

[Prime95]

Originally Posted by drkirkby

Is it possible to force a P-1 test on exponents one has a PRP test for?

It has a P-1 test, but with B1=1 000 000, B2=30 000 000. Is it possible to force another P-1 test, to higher values of B2, by exploiting my memory? If it is, the obvious next question to ask is whether it is sensible to do so or not.

Yes it is possible but you will be duplicating the B1=1000000 effort. Is it sensible? No.