Re: GETF Speedup?

On 11 Oct 2025, at 16:17, Yuri Davidovsky (as work at disclosure dot ie) <[email protected]> wrote:
> 
> Anyway, the lesson is: if you need fast key access, use lists, as strange as it sounds, just make sure they are built in a single sitting.

Here's LW's performance for a lookup miss with various structures of various sizes.

The way I did this was to build a table (plist, hash table, alist) containing entries from 0 to n-1, and then probe for n, averaging this over this a large number of times of course.  There is some loop overhead but it is identical for each type.

The behaviour of plists is interesting: I assume this is some cache in the machine.  Alists might do the same thing with a longer spine (but do not seem to).

Also below that is the same thing but for a hit at the half-way point, and first elt hit.

This is all an M1.

--tim