Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
A third compiler fully supports Fortran 2003!
29 views
Skip to first unread message
Damian Rouson
Nov 23, 2011, 12:12:10 AM11/23/11
to
For those interested, I just noticed that Portland Group has become
the third compiler vendor (after Cray and IBM) to offer complete
Fortran 2003 support and the first to do so on commodity/consumer
platforms, including Mac OS, Windows, and Linux.
Based on my testing of several other compilers, I believe 2012 will be
a turning point: I'd bet we end the year 2012 with more compilers
supporting Fortran 2003 than not.
Damian
the third compiler vendor (after Cray and IBM) to offer complete
Fortran 2003 support and the first to do so on commodity/consumer
platforms, including Mac OS, Windows, and Linux.
Based on my testing of several other compilers, I believe 2012 will be
a turning point: I'd bet we end the year 2012 with more compilers
supporting Fortran 2003 than not.
Damian
Richard Maine
Nov 23, 2011, 12:46:17 AM11/23/11
to
Damian Rouson <dam...@rouson.net> wrote:
> For those interested, I just noticed that Portland Group has become
> the third compiler vendor (after Cray and IBM) to offer complete
> Fortran 2003 support and the first to do so on commodity/consumer
> platforms, including Mac OS, Windows, and Linux.
That would be good news, but based on prior data, I have to ask...
> For those interested, I just noticed that Portland Group has become
> the third compiler vendor (after Cray and IBM) to offer complete
> Fortran 2003 support and the first to do so on commodity/consumer
> platforms, including Mac OS, Windows, and Linux.
Have you verified that the compiler actually shipping now does, in fact
fully support f2003?
The reason I ask is that I recall PGI having an announcement of their
f2003 compiler some time ago; it was mentioned in the newsgroup here.
But the announcement turned out to be a bit of an ...um... exageration.
I forget the details, but I recall that the compiler that was actually
shipping was missing major parts of f2003 and that the announcement
described more what they planned to do than what had actually been done.
Perhaps the plans have now come to fruition. But it also occurs to me
that you might be looking at the same misleading announcements I recall
from earlier. I find both possibilities believable. That's why I have to
ask whether you have verified what is actually shipping. The past
history makes me mistrust relying on the sales blurbs and docs on their
web site.
--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
Damian Rouson
Nov 23, 2011, 1:37:10 AM11/23/11
to
Richard Maine wrote:
> Damian Rouson<dam...@rouson.net> wrote:
>
>> For those interested, I just noticed that Portland Group has become
>> the third compiler vendor (after Cray and IBM) to offer complete
>> Fortran 2003 support and the first to do so on commodity/consumer
>> platforms, including Mac OS, Windows, and Linux.
>
> That would be good news, but based on prior data, I have to ask...
>
> Have you verified that the compiler actually shipping now does, in fact
> fully support f2003?
>
> The reason I ask is that I recall PGI having an announcement of their
> f2003 compiler some time ago; it was mentioned in the newsgroup here.
> But the announcement turned out to be a bit of an ...um... exageration.
> I forget the details, but I recall that the compiler that was actually
> shipping was missing major parts of f2003 and that the announcement
> described more what they planned to do than what had actually been done.
>
> Perhaps the plans have now come to fruition. But it also occurs to me
> that you might be looking at the same misleading announcements I recall
> from earlier. I find both possibilities believable. That's why I have to
> ask whether you have verified what is actually shipping. The past
> history makes me mistrust relying on the sales blurbs and docs on their
> web site.
>
I based this on the PG web site. I also just obtained a copy of the
> Damian Rouson<dam...@rouson.net> wrote:
>
>> For those interested, I just noticed that Portland Group has become
>> the third compiler vendor (after Cray and IBM) to offer complete
>> Fortran 2003 support and the first to do so on commodity/consumer
>> platforms, including Mac OS, Windows, and Linux.
>
> That would be good news, but based on prior data, I have to ask...
>
> Have you verified that the compiler actually shipping now does, in fact
> fully support f2003?
>
> The reason I ask is that I recall PGI having an announcement of their
> f2003 compiler some time ago; it was mentioned in the newsgroup here.
> But the announcement turned out to be a bit of an ...um... exageration.
> I forget the details, but I recall that the compiler that was actually
> shipping was missing major parts of f2003 and that the announcement
> described more what they planned to do than what had actually been done.
>
> Perhaps the plans have now come to fruition. But it also occurs to me
> that you might be looking at the same misleading announcements I recall
> from earlier. I find both possibilities believable. That's why I have to
> ask whether you have verified what is actually shipping. The past
> history makes me mistrust relying on the sales blurbs and docs on their
> web site.
>
compiler and will start testing it within a few days. Fingers crossed...
What I can say for sure is that there are other compilers for which I'm
certain the coverage of the Fortran 2003 standard will be quite broad in
2012 so that part of the post was accurate. I apologize for being
vague, but the most comprehensive and detailed assessment of compiler
support is of course the running survey that Chivers and Sleightholme
publish in ACM Fortran Forum. I feel confident predicting that the 2012
editions of their survey will show significant and broad progress across
multiple compilers.
Damian
Nasser M. Abbasi
Nov 23, 2011, 3:46:47 AM11/23/11
to
On 11/23/2011 12:37 AM, Damian Rouson wrote:
>>
> I based this on the PG web site. I also just obtained a copy of the
> compiler and will start testing it within a few days. Fingers crossed...
>
Does not Fortran have an official set of tests for compiler
>>
> I based this on the PG web site. I also just obtained a copy of the
> compiler and will start testing it within a few days. Fingers crossed...
>
conformity with the standard?
For example, Ada has that:
http://www.ada-auth.org/acats-files/3.0/docs/UG-1.HTM
"The Ada Conformity Assessment Test Suite (ACATS) provides
the official tests used to check conformity of an Ada implementation
with the Ada programming language standard (ANSI/ISO/IEC 8652:1995
and later corrigenda and amendments)"
If not, how else would one then verify that a given Fortran
compiler is conforming to a given standard such as F2003?
--Nasser
Ian Bush
Nov 23, 2011, 3:48:01 AM11/23/11
to
On 23/11/11 05:46, Richard Maine wrote:
> Damian Rouson<dam...@rouson.net> wrote:
>
>> For those interested, I just noticed that Portland Group has become
>> the third compiler vendor (after Cray and IBM) to offer complete
>> Fortran 2003 support and the first to do so on commodity/consumer
>> platforms, including Mac OS, Windows, and Linux.
>
> That would be good news, but based on prior data, I have to ask...
>
> Have you verified that the compiler actually shipping now does, in fact
> fully support f2003?
>
> The reason I ask is that I recall PGI having an announcement of their
> f2003 compiler some time ago; it was mentioned in the newsgroup here.
> But the announcement turned out to be a bit of an ...um... exageration.
> I forget the details, but I recall that the compiler that was actually
> shipping was missing major parts of f2003 and that the announcement
> described more what they planned to do than what had actually been done.
>
Well looking at the PG2011 Release Notes at
> Damian Rouson<dam...@rouson.net> wrote:
>
>> For those interested, I just noticed that Portland Group has become
>> the third compiler vendor (after Cray and IBM) to offer complete
>> Fortran 2003 support and the first to do so on commodity/consumer
>> platforms, including Mac OS, Windows, and Linux.
>
> That would be good news, but based on prior data, I have to ask...
>
> Have you verified that the compiler actually shipping now does, in fact
> fully support f2003?
>
> The reason I ask is that I recall PGI having an announcement of their
> f2003 compiler some time ago; it was mentioned in the newsgroup here.
> But the announcement turned out to be a bit of an ...um... exageration.
> I forget the details, but I recall that the compiler that was actually
> shipping was missing major parts of f2003 and that the announcement
> described more what they planned to do than what had actually been done.
>
http://www.pgroup.com/doc/pgirn.pdf
does look fairly comprehensive, page 19 onwards is probably the
interesting bit. And it does explicitly mention DTIO and parameterised
derived types which, I understand, are usually the stumbling blocks.
But I'm far from an expert here, so maybe somebody else could take a
look. And the release notes and reality might not quite match ...
Ian
William Clodius
Nov 23, 2011, 4:02:26 PM11/23/11
to
now NIST, had an official test suite for F77 that was never updated.
Dick Henderson sells a compiler test suite for later standards, but I
don't know if it has been updated for F2008.
--
Bill Clodius
los the lost and net the pet to email
Fritz Wuehler
Nov 24, 2011, 3:58:42 AM11/24/11
to
Ada *had* that. I don't believe there is a test after the Ada JPO was
closed. That means anything after '95 is a voluntary/self-claimed
complicance like FORTRAN is now.
> If not, how else would one then verify that a given Fortran
> compiler is conforming to a given standard such as F2003?
You cannot, since there is no test suite. However if you find a compiler
that claims conformance and pay for it, you can probably get the
manufacturer to fix it if you can prove it doesn't conform.
closed. That means anything after '95 is a voluntary/self-claimed
complicance like FORTRAN is now.
> If not, how else would one then verify that a given Fortran
> compiler is conforming to a given standard such as F2003?
that claims conformance and pay for it, you can probably get the
manufacturer to fix it if you can prove it doesn't conform.
0 new messages