mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   NFSNET Discussion (https://www.mersenneforum.org/forumdisplay.php?f=17)
-   -   Stopping on a dime? (https://www.mersenneforum.org/showthread.php?t=5093)

R.D. Silverman 2005-12-06 12:42

Stopping on a dime?
 
I am nearly done sieving 2,1402L.

Last night I did my first filtering pass on the data. The results
show (theoretically) that I could proceed with the LA if I wanted to....

I had a bare excess of 78 equations over ideals and a matrix of about
6.8 million rows.... I have never before seen such a small excess on the
first filter pass. Usually, I still have a defficiency after the first pass.

I will sieve for another two days. This will yield enough relations to
cut the matrix in half and thereby save about 3 weeks of Linear Algebra
time.....

I ran into a bug in my lattice siever when I did some initial experiments
on 2,781-. I was using a reciprocal quintic. This number is only 9 bits
bigger than 2,1402L, but I was only getting a tiny fraction of the
expected yield. I then tried 2,820+, (with a quartic), but this dumped
core. I am guessing that with the size of the factor bases I am using,
that polynomials lower than degree 6 are smooth so often that I am
running out of space to store some data. I need to investigate.
Unfortunately, I am swamped with real work, and don't have the time
right now. I have therefore switched to 2,1011+. I will do 2,781- at
a later time if someone else doesn't do it first.

Wacky 2006-03-20 23:46

Bob,
thank you for keeping us informed.


All times are UTC. The time now is 00:34.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.