View Single Post
Old 2021-09-02, 15:51   #507
VBCurtis's Avatar
Feb 2005
Riverside, CA

506910 Posts

Originally Posted by charybdis View Post
The best solution is to make CADO oversieve a bit to reduce the matrix size; you can do this by adding tasks.filter.required_excess = 0.05 to the params file (experiment with different values to see what's best for your system).
I use 0.05 for C140-C160 to reduce matrix time when running all jobs on a single machine; I think EdH (who sieves on a "farm" of a dozen or more smaller machines) uses 0.08 or 0.10. The number is a percentage of total relations, so 0.05 asks CADO to make sure there are 5% more relations than strictly necessary for filtering; this "extra" yields a usefully smaller matrix. There are diminishing returns- do not expect a setting of 0.20 to make a 30-minute matrix!

A setting of 0.04 or 0.05 also reliably (?) produces enough relations for msieve filtering to work, useful if you're using that combination of CADO sievers and msieve postprocessing.
VBCurtis is online now   Reply With Quote