Thread: yafu bugs
View Single Post
Old 2019-01-12, 03:52   #470
bsquared
 
bsquared's Avatar
 
"Ben"
Feb 2007

22×23×37 Posts
Default

I should maybe also (re-)mention that if you don't want to search for a fixed time, you can use the -psearch option with either min,avg, or good as a qualifier. yafu will stop when it finds a {min,avg,good} poly for the input number.

min,avg,good are chosen based on Batalov's heuristic.

-psearch min --> uses a multiplier of 1 to that fit
-psearch avg --> uses a multipler of 1.036
-psearch good --> uses a multiplier of 1.072

Just now, with that c120, I found a "min" poly of score = 2.937000e-010 in 3 minutes (8 threads), after searching a range of only a thousand coefficients or so. That actually just barely missed being qualified "avg".

On the start of the job the heuristic said the following, so the poly fits in the range expected:
nfs: expecting degree 5 poly E from 2.87e-010 to > 3.30e-010

best poly:
# norm 2.039066e-011 alpha -6.282017 e 2.937e-010 rroots 5
n: 227010481295437363334259960947493668895875336466084780038173258247009162675779735389791151574049166747880487470296548479
skew: 95431.04
c0: -46602088970638410228907101375
c1: 643792960629014034531255
c2: 64059317395932803683
c3: 49937622929009
c4: -7589830536
c5: 9108
Y0: -120039626008868252603764
Y1: 93171618323

Last fiddled with by bsquared on 2019-01-12 at 04:20
bsquared is offline   Reply With Quote