caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Jon Harrop <jon@ffconsultancy.com>
To: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] HLVM stuff
Date: Mon, 28 Sep 2009 02:25:44 +0100	[thread overview]
Message-ID: <200909280225.44292.jon@ffconsultancy.com> (raw)
In-Reply-To: <4DB7BEFB-70C6-4ADA-B0F9-4ED0717A85B6@refined-audiometrics.com>

On Monday 28 September 2009 01:35:32 David McClain wrote:
> Yes, this is beginning to sound very interesting... So now that you
> have F#, which I understand to be some derivative of OCaml,

F# is superficially similar to OCaml, most notably its OCaml-like syntax, but 
there are some quite major differences:

http://www.strangelights.com/fsharp/wiki/default.aspx/FSharpWiki/FSharpAndOCaml.html

> why do you need HLVM?

Good question. I saw these important advantages realized in F# by Microsoft 
and wanted to bring those benefits to the OCaml/Linux world. There is 
no "need" to do so unless you refuse to use Windows and I am happily using 
Windows now. Moreover, the libraries available under Linux are dire in 
comparison to .NET. Hence I am no longer really motivated to work on HLVM. F# 
is a lot more fun and a lot more profitable. :-)

> Is F# using the LLVM?

No. F# is Microsoft's new programming language for .NET.

> or is it executing natively compiled code?

Yes. The F# compiler generates .NET assemblies containing CIL (Common 
Intermediate Language) that the CLR (Common Language Run-time) then JIT 
compiles the CIL to native code:

  http://en.wikipedia.org/wiki/File:CLR_diag.svg

This is true of the interactive F# REPL as well as compiled binaries.

>  From what I have garnered today in a quick scan of JIT docs, it
> appears that JIT cannot compete yet with native code. But if the
> timings you stated are for some kind of JIT against byte-codes, I am
> very impressed.

The timings I posted show JIT-compiled F# solving your problem orders of 
magnitude faster than native-code compiled with ocamlopt. OCaml's interpreted 
bytecode is even slower than its compiled native code, of course. I don't 
know how fast other native-code compiled languages like C, C++ and Fortran 
are in comparison except that some of my numerical F# code outperform's 
Intel's vendor-tuned Fortran running on Intel hardware.

-- 
Dr Jon Harrop, Flying Frog Consultancy Ltd.
http://www.ffconsultancy.com/?e


  reply	other threads:[~2009-09-28  1:25 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-09-27 17:43 David McClain
2009-09-27 19:25 ` [Caml-list] " Jon Harrop
2009-09-27 21:58   ` David McClain
2009-09-27 23:14     ` Jon Harrop
2009-09-28  0:35       ` David McClain
2009-09-28  1:25         ` Jon Harrop [this message]
2009-10-13 22:18         ` Jon Harrop

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200909280225.44292.jon@ffconsultancy.com \
    --to=jon@ffconsultancy.com \
    --cc=caml-list@yquem.inria.fr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).