For a tessellation function, I believe I should use a hash table or a alist to
save the index of some points to prevent duplicates. Yesterday I felt I should
test how fast would the hash-table work to index my 3d coordinates. For that
reason, I wrote that small benchmark and realized that the destructive function
on alist gives impressive speed agains the non destructive loop. The hash table
gives results that are close to the alist (destructive).
To use the alist, I’d have to implement a hashing function which could then
give results pretty similar to the hash table. But the non destructive alist
scares me a little.
To do the same job, it takes 650s for the non destructive function and 0.014s
for the use of destructive “alist-update!”. Anyone can explain why is this so
slow?
http://paste2.org/ng2fbFyk
Also what is the big difference between alist and a hash-table other than alist
doesn’t seem to hash objects. After testing again, I realize that
“alist-update!” isn’t even adding new element to the current list.
--
Loïc Faure-Lacroix
_______________________________________________
Chicken-users mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/chicken-users
