[racket] Question about hash table efficiency

From: Sam Tobin-Hochstadt (samth at ccs.neu.edu)
Date: Thu Feb 14 14:18:53 EST 2013

On Thu, Feb 14, 2013 at 2:15 PM, Robby Findler
<robby at eecs.northwestern.edu> wrote:
> Is this a bottleneck in some real program? Perhaps you'd share that (or some
> inner loop of it)? I wonder if there is a better way to speed it up than
> this.

`hash-set!` (which, like `hash-update`, is written in Racket and
hashes twice) is a real bottleneck in Typed Racket, in particular the
duplicated hash computation. I believe this is mostly because of
interning, where `equal?` hashes are used.

A version of `hash-update` which hashed only once could (probably)
speed this up significantly.


Posted on the users mailing list.