[Caml-list] Inline assembly?

[Caml-list] Inline assembly?

[Niall Dalton]

Hi there,

Is it possible to use inline assembly in ocaml?
If not, what is the lowest overhead way to approximate
it? Using a C function which includes the asm?

Best regards,
Niall
-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr

Re: [Caml-list] Inline assembly?

[Xavier Leroy]

> Is it possible to use inline assembly in ocaml?

Currently, no.  A gcc-style inline asm facility could be useful to
express some canned instruction sequences that are currently hardcoded
into each native-code generator, but that's also quite a lot of work
to implement properly.

> If not, what is the lowest overhead way to approximate
> it? Using a C function which includes the asm?

Yes.  Since the C function will not heap-allocate nor raise
exceptions, you can tell the compiler about it using the "noalloc"
annotation:

        external myfunction : ... = "myfunction" "noalloc"

and this will select a faster calling convention from Caml to C for
this function.  If the function operates over floats, there are other
performance tricks one can play; write me if this is the case.

Regards,

- Xavier Leroy
-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr

Re: [Caml-list] Inline assembly?

[Markus Mottl]

On Thu, 22 Mar 2001, Xavier Leroy wrote:
> Yes.  Since the C function will not heap-allocate nor raise
> exceptions, you can tell the compiler about it using the "noalloc"
> annotation:
> 
>         external myfunction : ... = "myfunction" "noalloc"
> 
> and this will select a faster calling convention from Caml to C for
> this function.  If the function operates over floats, there are other
> performance tricks one can play; write me if this is the case.

I have some functions that return floats in the LAPACK-interface. They
could (slightly) benefit from more appropriate calling conventions.

If I am not mistaken, it is not ok to use "noalloc" if the C-function uses
"copy_double", because the latter could trigger the GC. If the function
returns a float, it seems that the additional annotation "float" could
be applied, but it is obviously not always used in the Pervasives library
(e.g. float_of_string; ld_exp).

Is this just an oversight or do these functions have other properties that
prevent the use of this annotation? Are there special rules to follow?

Best regards,
Markus Mottl

-- 
Markus Mottl, mottl@miss.wu-wien.ac.at, http://miss.wu-wien.ac.at/~mottl
-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr

Re: [Caml-list] Inline assembly?

[Xavier Leroy]

> > If the function operates over floats, there are other
> > performance tricks one can play; write me if this is the case.
> 
> I have some functions that return floats in the LAPACK-interface. They
> could (slightly) benefit from more appropriate calling conventions.
> 
> If I am not mistaken, it is not ok to use "noalloc" if the C-function uses
> "copy_double", because the latter could trigger the GC.

Correct.

> If the function
> returns a float, it seems that the additional annotation "float" could
> be applied, but it is obviously not always used in the Pervasives library
> (e.g. float_of_string; ld_exp).

Currently, the "float" annotation works only for external functions
where the result and all parameters are of type "float".  Typically,
transcendental and trigonometric functions.  In this case, the
arguments and result are passed unboxed, as C "double" floats.  But it
doesn't work for functions that take / return a mixture of floats and
other data types.

- Xavier Leroy
-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr