Intel® Fortran Compiler for Linux* and Mac OS X*

Calling-convention extensions for F2003?

It would be nice if ISO_C_BINDING could replace the !DEC$ directives involved in C binding. One obvious missing feature is the ability to define a calling convention other than the standard cdecl. This is mainly an issue for Win32 platforms, because it does not follow ISO C conventions. It also seems that Fortran Standards members do not have a big interest in adding non-ISO-C support.

intel fortran compiler version 7.1

I m searching for downloadable ifc version 7.1 in intel website but i can't able to find it instead of ifc version 9.0. I tried to compile VASP 4.6 with ifc9.0 and i have succesffuly compiled it but there was error of segmentation fault when running it. And i also tried compiled with ifc8.1 and i got the error of 'routine ZPOTRF failed'. Some people suggest to use ifc7.1 to solve these problem of VASP. How can i download the non-commercial ifc7.1?

error while loading shared libraries: libimf.so:

Hi.
I cannot complie my program using the makefile. However, if I type the complie command, it can complie. I donot know why this situation happens.

sms is executable file after linked. This messege is below after I run.

[root@tmp code]# ./sms
./sms: error while loading shared libraries: libimf.so: cannot open shared object file: No such file or directory ----> What do I do?

This is manual but I do not know what I do. I think this is some clue for using the makefile.

New feature?

I'd like to get a report from the Intel FORTRAN compiler like the one that the IRIX64 MIPSpro FORTRAN compiler would generate when using

-FLIST:=ON

Basically, I want to see what the source code looks like after all the optimization (including interprocedural) has been done.

ifort does not seem to have this feature. I've played with the -opt_report options, but the reports aren't clear to me. Do you know of any undocumented options that would help me get optimized source code from the compiler?

Stack limit question

Hi,
I would like to ask you a small question concerning the ifort9.0 version which i'm using under lunix to compile my code, I read a replay from luxiaod user who had a problem of stack size, my problem is similar, my code works well up to some size of arrays, but it crashs with the message " erreur de segmentation" when I increase this array, I would like to know how luxiaod did to increase this stack size?
thanks for all.

Message Edited by Steve_Lionel on 05-10-2006 03:08 PM

Cluster OpenMP - more information?

Hello,

on the compiler product pages of icc 9.1 and ifort 9.1 I found that they now support to use (slightly modified?) OpenMP directives to compile/run a program on a distributed-memory computer cluster.

Which is termed "Cluster OpenMP".

Unfortunally, the web is rather oblique on the details.

One finds a whitepaper:

http://cache-www.intel.com/cd/00/00/28/58/285865_285865.pdf

A conference of a high-performance computation center:

Intel Fortran Compiler 9.1 is Now Available

Intel Fortran 9.1 for Linux* is now available. New features include:

- Improved run-time performance optimizations
- Support for additional features from Fortran 2003
- Optimizations to enhance performance on the new Intel Core and Intel Core 2 processors

Please see the product release notes for further details.

Customers can download the new version from Intel Premier Support at no additional charge. (Downloads through Intel Premier Support will be available later today.) FTP downloads are available through the following link:

订阅 Intel® Fortran Compiler for Linux* and Mac OS X*