Fortran - Re: Linux 64-bit (Opteron) Fortran 95?

Author

Re: Linux 64-bit (Opteron) Fortran 95?

Toon Moene

2004-03-27, 12:16 am

Tim Prince wrote:

> If you're talking about x86-64, maybe you could say there is no "purely
> 64-bit Linux" for that hardware. You could say that for any 64-bit system
> which has even rudimentary 32-bit support.

Debian x86-64 "pure64" is in the making - it is by no means a finished
product yet, however.

--
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)

Scott Robert Ladd

2004-03-27, 12:17 am

On Thu, 18 Mar 2004 00:46:31 +0000, Tim Prince wrote:
> I'm still mystified about your point. If you have installed the full
> gcc/g77 system on x86-64 linux, you can build and run both 32- and 64-bit
> binaries. You can install and use commercial 32-bit compilers, if you can
> figure out how to cause them to stick with the 32-bit software components.
> In my experience, that has required building a 32-bit copy of mpi, for
> applications requiring that.

The glibc (as well as everything else in the system lib directories) on my
gentoo-amd64 system is 64-bit; therefore, it does me no good to compile
code for 32-bit, since it won't link with the system C library.
Gentoo-amd64 does not install 32-bit libraries on AMD64; I could, I
suppose, set up a chroot with 32-bit libraries, but I don't see the point.

> If you're talking about x86-64, maybe you could say there is no "purely
> 64-bit Linux" for that hardware. You could say that for any 64-bit system
> which has even rudimentary 32-bit support.

The Linux kernel can be compiled without 32-bit executable support. While
I built my kernel with 32-bit capabilities, the lack of any 32-bit
libraries makes 32-bit programs useless, unless they are statically linked.

Debian pure64 will also be a 100% 64-bit installation.

--
Scott Robert Ladd
Coyote Gulch Productions (http://www.coyotegulch.com)
Software Invention for High-Performance Computing

Hanyou Chu

2004-03-27, 12:17 am

I don't know if it is valid code. I posted another message
asking the question, but it didn't show up. Maybe I should ask
here. It's something like this

subroutine callee(n,array)
integer n
real(8) array(n)
....
end

subroutine caller()
type the_type
real(8) x,y
end type
type(the_type) a(10)
...
call callee(10,a.x)
...
end

They are both in the same file. I assume that CVF and Intel recognize
this, automatically use temporary arrays to handle it. The code
seem to work fine. Is it valid according to the standard?

It was written by a colleague of mine. I wouldn't write
something like that myself.

>
> It would be much more interesting to know whether it refused to
> compile code that was valid.
>
> Perhaps the code in question is valid, but just that some other
> compilers accepted it is not very conclusive evidence of that.
> It can be of some supporting value, but doesn't stand well on
> it's own.
>
> Is the code independently believed to be valid, or is the fact
> that it compiled with some compilers just assumed to imply such?
>
> Specifics? I.E. Is it not accepting some particular extension
> accepted by the other compilers in question? If so, that would
> be a lot more constructive to know than just that there is some
> code of some unspecified nature that it doesn't accept.
>
> I'm not "defending" or otherwise commenting on the Absoft compiler;
> I don't know enough about it to do so. I'm just tying to encourage
> more informative data about what you are finding.

Udo Grabowski

2004-03-27, 12:17 am

Hanyou Chu wrote:
> call callee(10,a.x)
"a.x" is not a valid Fortran statement. The correct syntax
is "a%x". If that has been accepted by some compilers, it
was done so by a vendor specific language extension.
--
Dr. Udo Grabowski email: udo.grabowski@imk.fzk.de
Institut f. Meteorologie und Klimaforschung ASF,Forschungszentrum Karlsruhe
Postfach 3640, 76021 Karlsruhe, Germany Tel: (+49) 7247 82-6026
http://www.fzk.de/imk/asf/ame/grabowski/ Fax: " -7026

Sponsored Links