Re: [PATCH] Re: Bignum performance

[Emanuel Berg]

Ihor Radchenko wrote:

> You can again compare Elisp with CL and let us know what is
> being noticeably slower. It will be an indication that
> something might be improved.

Her is yet another file,

  https://dataswamp.org/~incal/cl/bench/inclist.cl

The Elisp time for this benchmark is, with `native-comp-speed'
at the default - for the benchmarks package - maximal
optimization level, namely 3 - at 5.86 sec, while the CL is at
1.635993 sec.

So here, CL is 258% faster.

I run the benchmarks from Emacs, and the CL also from Emacs,
with SLIME and SBCL. So that should be pretty fair, but maybe
when all benchmarks are translated one could do the Elisp with
batch and the SBCL not using Emacs at all.

-- 
underground experts united
https://dataswamp.org/~incal