From: Martin Uecker <muecker@gwdg.de>
To: Tobias Burnus <tburnus@baylibre.com>,
Janne Blomqvist <blomqvist.janne@gmail.com>,
FX Coudert <fxcoudert@gmail.com>
Cc: Thomas Koenig <tkoenig@netcologne.de>,
Mark Wielaard <mark@klomp.org>, <overseers@sourceware.org>,
<gcc@gcc.gnu.org>, <binutils@sourceware.org>,
<gdb@sourceware.org>, <libc-alpha@sourceware.org>,
"fortran@gcc.gnu.org" <fortran@gcc.gnu.org>
Subject: Re: Generated files in libgfortran for Fortran intrinsic procedures (was: Updated Sourceware infrastructure plans)
Date: Thu, 18 Apr 2024 14:32:16 +0200 [thread overview]
Message-ID: <56b27622e866daa5871afa1239866d47f23609dc.camel@gwdg.de> (raw)
In-Reply-To: <686fac37-698f-4373-9469-ccabe96e702a@baylibre.com>
Am Donnerstag, dem 18.04.2024 um 14:01 +0200 schrieb Tobias Burnus:
> Hi Janne,
>
> Janne Blomqvist wrote:
> > back when I was active I did think about this
> > issue. IMHO the best of my ideas was to convert these into C++
> > templates.
I haven't looked at libgfortran but I didn't find it problematic
at all to use C in similar numerical code and this helps
with portability.
Either I use macros, which I keep short and then do not find
inferior to templates (having used C++ for years previously) or
- if there is really a lot of code that needs to be specialized
for a type - simply by using includes:
#define matmul_type double
#include "matmul_impl.c"
Martin
>
> I think this will work – but we have to be super careful:
>
> With C++, there is the problem that we definitely do not want to add
> dependency on libstdc++ nor to use some features which require special
> hardware support (like exceptions [always bad], symbol aliases, ...). —
> On some systems, a full C++ support might be not available, like
> embedded systems (including some odd embedded OS) or offloading devices.
>
> The libstdc++ dependency would be detected by linking as we currently
> do. For in-language features, we have to ensure the appropriate flags
> -fno-exceptions (and probably a few more). And it should be clear what
> language features to use.
>
> If we do, I think that would surely be an option.
>
> > What we're essentially doing with the M4 stuff and the
> > proposed in-house Python reimplementation is to make up for lack of
> > monomorphization in plain old C. Rather than doing some DIY templates,
> > switch the implementation language to something which has that feature
> > built-in, in this case C++. No need to convert the entire libgfortran
> > to C++ if you don't want to, just those objects that are generated
> > from the M4 templates. Something like
> >
> > template<typename T>
> > void matmul(T* a, T* b, T* c, ...)
> > {
> > // actual matmul code here
> > }
> >
> > extern "C" {
> > // Instantiate template for every type and export the symbol
> > void matmul_r4(gfc_array_r4* a, gfc_array_r4* b, gfc_array_r4* c, ...)
> > {
> > matmul(a, b, c, ...);
> > }
> > // And so on for other types
> > }
>
> Cheers,
>
> Tobias
next prev parent reply other threads:[~2024-04-18 12:32 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-17 23:27 Updated Sourceware infrastructure plans Mark Wielaard
2024-04-18 6:04 ` Thomas Koenig
2024-04-18 8:14 ` FX Coudert
2024-04-18 9:01 ` Christophe Lyon
2024-04-18 11:38 ` Janne Blomqvist
2024-04-18 12:01 ` Generated files in libgfortran for Fortran intrinsic procedures (was: Updated Sourceware infrastructure plans) Tobias Burnus
2024-04-18 12:32 ` Martin Uecker [this message]
2024-04-19 9:35 ` Updated Sourceware infrastructure plans Jonathan Wakely
2024-04-18 15:56 ` Joseph Myers
2024-04-18 17:37 ` Frank Ch. Eigler
2024-04-18 17:54 ` Joseph Myers
2024-04-18 18:29 ` Matt Rice
2024-04-22 15:39 ` Tom Tromey
2024-04-23 2:55 ` Jason Merrill
2024-04-23 3:12 ` Simon Marchi
2024-04-23 3:24 ` Tom Tromey
2024-04-23 3:51 ` Jason Merrill
2024-04-23 8:56 ` Mark Wielaard
2024-04-23 9:39 ` Richard Earnshaw (lists)
2024-04-23 15:08 ` Tom Tromey
2024-04-23 15:25 ` Simon Marchi
2024-04-24 8:49 ` Aktemur, Tankut Baris
2024-04-23 4:06 ` Ian Lance Taylor
2024-04-23 9:30 ` Richard Earnshaw (lists)
2024-04-23 13:51 ` Ian Lance Taylor
2024-05-01 19:15 ` Jeff Law
2024-05-01 19:38 ` Jonathan Wakely
2024-05-01 20:20 ` Mark Wielaard
2024-05-01 20:53 ` Tom Tromey
2024-05-01 21:04 ` Simon Marchi
2024-05-02 15:35 ` Pedro Alves
2024-05-02 23:05 ` Fangrui Song
[not found] ` <DS7PR12MB57651DA3A5C22B2847C13580CB182@DS7PR12MB5765.namprd12.prod.outlook.com>
2024-05-07 16:17 ` Joseph Myers
2024-05-10 10:43 ` Ben Boeckel
2024-05-01 20:04 ` Jason Merrill
2024-05-01 21:26 ` Mark Wielaard
2024-05-01 22:01 ` Sergio Durigan Junior
2024-05-02 12:54 ` Claudio Bantaloukas
2024-05-02 15:33 ` Pedro Alves
2024-05-03 2:59 ` Ian Lance Taylor
2024-05-04 19:56 ` Ben Boeckel
2024-05-05 5:22 ` Benson Muite
2024-05-06 13:58 ` Ben Boeckel
2024-05-07 16:26 ` Joseph Myers
2024-05-01 21:38 ` Jeff Law
2024-05-02 6:47 ` Richard Biener
2024-05-02 11:29 ` Ian Lance Taylor
2024-05-02 14:26 ` Simon Marchi
2024-05-02 11:45 ` Mark Wielaard
2024-05-01 22:56 ` Tom Tromey
2024-04-23 10:34 ` Florian Weimer
2024-04-22 10:01 ` Mark Wielaard
2024-04-22 13:23 ` Joseph Myers
2024-04-19 9:33 ` Jonathan Wakely
2024-04-22 10:24 ` Mark Wielaard
2024-04-22 11:40 ` Jonathan Wakely
2024-04-23 0:48 ` Frank Ch. Eigler
2024-05-16 15:58 ` Cristian Rodríguez
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=56b27622e866daa5871afa1239866d47f23609dc.camel@gwdg.de \
--to=muecker@gwdg.de \
--cc=binutils@sourceware.org \
--cc=blomqvist.janne@gmail.com \
--cc=fortran@gcc.gnu.org \
--cc=fxcoudert@gmail.com \
--cc=gcc@gcc.gnu.org \
--cc=gdb@sourceware.org \
--cc=libc-alpha@sourceware.org \
--cc=mark@klomp.org \
--cc=overseers@sourceware.org \
--cc=tburnus@baylibre.com \
--cc=tkoenig@netcologne.de \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).