Re: [Caml-list] Efficient OCaml multicore -- roadmap?

[Hugo Ferreira <hmf@inescporto.pt>]

Fabrice,

Don't mean to stoke this fire but....

On 04/21/2011 08:48 AM, Fabrice Le Fessant wrote:
> On 04/21/2011 01:23 AM, Jon Harrop wrote:
>> The OCaml team at INRIA are not motivated to do this because it does not
>> constitute research, would probably make Coq slower and would burden them
>> with maintaining irrelevant features.
>
> You think the programmers in the world that are only interested in
> floating-point intensive computations, with fine-grain concurrency, are
> a huge majority. I think they are not so many. Can we do a better job of
> quantifying this ?
>

Some of us are interested in parallel computation for symbolic
manipulation. In my case I am manipulating sets of literals in
predicate logic. I am sure this is also useful in other areas
such ans ILP, graph mining, frequent pattern mining, etc.

>> OCaml users just migrate to other languages that are closer to what they
>> want rather than spending years learning how to build a parallel OCaml, then
>> doing it and then building a community around it. The only notable exception
>> might be Jane St. Capital because they have the resources and a vested
>> interest in performance. There are other large companies invested in OCaml,
>> like Citrix, but they aren't so interested in parallelism.
>
> That's probably the reason why so many scientists use Python instead of
> OCaml, because it is faster with better multicore support ?  I always
> thought Python was slower than OCaml, and had no multicore support...
>

Maybe the issue here is not why people use Python but why more people
don't use Haskell, ML, Ocaml, F#, etc.

Regards,
Hugo F.


> Fabrice
>