[Agda] Re: Fwd: Hashtables: 'Basic' hashtables performance quirks on Win64

[Andreas Abel]

Kyra: Note there is an issue tracker for hashtables at:

   https://github.com/gregorycollins/hashtables/issues

On 18.01.14 12:25 PM, Andreas Abel wrote:
> Found this message on libraries at haskell.org...
>
> I guess Agda is a serious test case for hashing methods.  Last year, I
> had problems with a bad version of the 'hashable' package.
>
> Kyra, none of the main Agda developers uses Windows.  You could help by
> making a patch that conditionally (#if) chooses a hash table
> implemenation that works for your case.
>
> Cheers,
> Andreas
>
> -------- Original Message --------
> Subject: Hashtables: 'Basic' hashtables performance quirks on Win64
> Date: Sat, 18 Jan 2014 13:38:34 +0400
> From: kyra <kyrab at mail.ru>
> To: Haskell Libraries <libraries at haskell.org>
>
> Having built Agda on 64-bit GHC 7.6.3 on Windows I've stumbled on Agda
> being unbearable slow (I tried it on Penryn and Sandy Bridge computers).
> The culprit turned out to be Agda uses 'BasicHashTable' in
> src/full/Agda/TypeChecking/Serialise.hs (line 98).
>
> Switching either to Cuckoo or Linear immediately cures the problem.
>
> The problem does not manifest itself on 32-bit Windows build or any
> Linux build.
>
> Also 'Basic' hashtable perform better than Cuckoo and Linear in standard
> benchmarks on Win64. And at the same time in Agda scenario it's
> performance degrades by a *couple of orders of magnitude*!
>
> Thus, *all* things are important here: 64-bit, Windows, Agda use case.
>
> I wonder what stands behind this behaviour? What are specifics of
> 'Basic' hashtable implementation and how to fix this or should this be
> fixed at all?
>
> Regards,
> Kyra
> _______________________________________________
> Libraries mailing list
> Libraries at haskell.org
> http://www.haskell.org/mailman/listinfo/libraries
>
>
>
>