Hmm... I guess then what I was seeing was the consequence of hash table collisions -- I imagine these become less frequent as the table grows.
Thanks, -- Raul On Tue, Mar 1, 2022 at 3:54 PM Henry Rich <henryhr...@gmail.com> wrote: > > The code is in a.c . The hashing needs to be rewritten, but the code is > simple, just a hash of the argument and a corresponding table of values. > > The cacheability test is in int0 at a.c:143. It appears to cache any > atomic [x/]y that is tolerantly equal to an integer. > > Henry Rich > > On 3/1/2022 3:43 PM, Raul Miller wrote: > > P.S. further testing suggests that M. will eventually cache for larger > > values of y. But not initially. Which is interesting... I wish I > > understood the mechanism better. > > > > > -- > This email has been checked for viruses by AVG. > https://www.avg.com > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
