mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Math (https://www.mersenneforum.org/forumdisplay.php?f=8)
-   -   Time to prp equation (https://www.mersenneforum.org/showthread.php?t=3830)

jasong 2005-03-10 23:28

Time to prp equation
 
Hey, guys, I would like to sieve until the sieving rate is JUUUUUUUUST under the average prp rate. Assuming that the minimum and maximum exponent are known and they increase at a steady rate(is this a good assumption?) how do I figure out the average length of time for a test?

Btw, I've sieved these already, but have no idea if I've reached the goal stated above. :help:

rogue 2005-03-11 00:49

[QUOTE=jasong]Hey, guys, I would like to sieve until the sieving rate is JUUUUUUUUST under the average prp rate. Assuming that the minimum and maximum exponent are known and they increase at a steady rate(is this a good assumption?) how do I figure out the average length of time for a test?

Btw, I've sieved these already, but have no idea if I've reached the goal stated above. :help:[/QUOTE]

I assume you are using PRP, LLR or PFGW. It should be fairly easy. This is what I do (assuming fixed k as opposed to fixed n). Here is an example from which you can extrapolate:

low k = 50000
high k = 200000

Calculate midk where midk = 200000 - (200000 - 50000)/4. I choose midk like this because it will take 16 times longer to PRP test k=200000 than k=50000. midk is about 3/4 of the range about lowk.

Run PRP/LLR/PFGW on midk*base^n+/-1 doing about 10% of the PRP test.

Multiply the time it took to do 10% by 10.

Use that time to be the breakpoint to change over from sieving to PRP testing.

I'm certain that others have different methods.

Use that time

jasong 2005-03-11 03:30

[QUOTE=rogue]I assume you are using PRP, LLR or PFGW. It should be fairly easy. This is what I do (assuming fixed k as opposed to fixed n). Here is an example from which you can extrapolate:

low k = 50000
high k = 200000

Calculate midk where midk = 200000 - (200000 - 50000)/4. I choose midk like this because it will take 16 times longer to PRP test k=200000 than k=50000. midk is about 3/4 of the range about lowk.

Run PRP/LLR/PFGW on midk*base^n+/-1 doing about 10% of the PRP test.

Multiply the time it took to do 10% by 10.

Use that time to be the breakpoint to change over from sieving to PRP testing.

I'm certain that others have different methods.

Use that time[/QUOTE]I'm still trying to figure out these instructions(I'm sure they make sense, it's just that I haven't done anything but simple math for awhile).

Thanks, though.


All times are UTC. The time now is 17:00.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.