[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: Re: LuaJIT missing from language shootout benchmarking site
- From: Isaac Gouy <igouy2@...>
- Date: Tue, 17 May 2011 14:53:04 -0700 (PDT)
--- On Tue, 5/17/11, Francesco Abbate <firstname.lastname@example.org> wrote:
> > Was inclusion of beta LuaJIT for the last 5 years very
> fortunate? ;-)
> Why you should think it was not fortunate ?
I did not say inclusion of beta LuaJIT was not fortunate.
I asked Leo Razoumov whether inclusion of beta LuaJIT was fortunate.
> The inclusion of LuaJIT was actually very fortunate ...
Yes, for 5 years the benchmarks game publicised LuaJIT beyond the Lua community.
> >> Well, arbitrary decisions are what they are --
> arbitrary decisions
> > Are you talking about the previous inclusion of
> LuaJIT? :-)
> and what about ATS, the programming language that nobody
> uses, that
> does not even have a real compiler (it does generate C
> code) and with
> all the samples of code written by a single person (the
> maintainer of
> the ATS compiler) ?
How many of the LuaJIT code samples do you think were written by a single person, the maintainer of LuaJIT? :-)
What about ATS? - "To show working programs written in less familiar programming languages"
> Why the custom memory allocator was rejected for the binary
> tree application ? That was proposed many times for C and it
> seems that there was no logical explication for the ban.
Do we need to see 100 different C custom memory allocators before we understand that a custom implementation will likely be faster than a more general implementation?
Rather than 100 different custom memory allocators designed specially for a tiny piece of benchmark code, I think it would be more interesting to see the performance of a general purpose implementation like tcmalloc.
> Yes, the choices of the computer benchmark game are pretty
The distinction isn't between arbitrary choices and principled choices - it's between choices you like and choices you don't like :-)