Intel Fortran 2003 Status
in [Fortran]
|
From: Jason Blevins on 11 Aug 2010 10:50 The Fortran Forum's status table for Fortran 2003 support for Intel seems to be somewhat dated. According to Intel's article on Fortran Language Standards [1] from June 2009, the following are supported in version 11.1 but are still listed as not supported in the table: * Deferred bindings and abstract types * Type-bound procedures * ABSTRACT attribute * DEFERRED attribute I've verified this by compiling some simple example programs using these features. So, it seems to me that the following rows in the Intel column should have a "Y" instead of an "N": * Procedures bound by name to a type * Deferred bindings and abstract types I'm happy to contact Ian (if he's not reading this already) but I wanted to get some independent verification first, just in case I'm misunderstanding something. Perhaps someone knows of other rows that should be updated as well? [1]: http://sdf.org/l/c43 -- Jason R. Blevins Assistant Professor of Economics Ohio State University http://jblevins.org/
From: Paul van Delst on 11 Aug 2010 11:31 From the "in depth" info for ifort v11.1: http://software.intel.com/sites/products/collateral/hpc/compilers/flin_indepth.pdf which I found searching for "ifort f2003" <quote> More Fortran 2003 Support: Additional support includes object-oriented features such as CLASS declaration, SELECT TYPE constant, inheritance association, and more. Other additions support deferred-length character entities, PUBLIC types with PRIVATE components (and vice versa), ENUMERATOR, IEEE Floating Point Exception Handling, ALLOCATE extensions, array constructor changes, and more to bring your Fortran apps closer to the standard. These join C interoperability features introduced in the last release to make it easier to develop mixed-language applications. </quote> I guess the features in question fall into the "and more" category? Jason Blevins wrote: > The Fortran Forum's status table for Fortran 2003 support for Intel > seems to be somewhat dated. According to Intel's article on Fortran > Language Standards [1] from June 2009, the following are supported in > version 11.1 but are still listed as not supported in the table: > > * Deferred bindings and abstract types > * Type-bound procedures > * ABSTRACT attribute > * DEFERRED attribute > > I've verified this by compiling some simple example programs using > these features. So, it seems to me that the following rows in the > Intel column should have a "Y" instead of an "N": > > * Procedures bound by name to a type > * Deferred bindings and abstract types > > I'm happy to contact Ian (if he's not reading this already) but I > wanted to get some independent verification first, just in case I'm > misunderstanding something. Perhaps someone knows of other rows that > should be updated as well? > > [1]: http://sdf.org/l/c43 >
From: Ian on 12 Aug 2010 07:18 On Aug 11, 3:50 pm, Jason Blevins <jrble...(a)sdf.org> wrote: > The Fortran Forum's status table for Fortran 2003 support for Intel > seems to be somewhat dated. According to Intel's article on Fortran > Language Standards [1] from June 2009, the following are supported in > version 11.1 but are still listed as not supported in the table: > > * Deferred bindings and abstract types > * Type-bound procedures > * ABSTRACT attribute > * DEFERRED attribute > > I've verified this by compiling some simple example programs using > these features. So, it seems to me that the following rows in the > Intel column should have a "Y" instead of an "N": > > * Procedures bound by name to a type > * Deferred bindings and abstract types > > I'm happy to contact Ian (if he's not reading this already) but I > wanted to get some independent verification first, just in case I'm > misunderstanding something. Perhaps someone knows of other rows that > should be updated as well? > > [1]:http://sdf.org/l/c43 > > -- > Jason R. Blevins > Assistant Professor of Economics > Ohio State Universityhttp://jblevins.org/ Thanks for bringing this up. > * Procedures bound by name to a type should be partial. I'll correct in the next edition. > * Deferred bindings and abstract types The information I use for the Intel entry is provided by one of the compiler team at Intel, Stan Whitlock. I should have changed no to partial in the entry above, but missed it in the last update. The other is still no as far as I can determine. I'm on holiday at the moment and don't have complete access to my records. Most of the information in the table comes from compiler people rather than their web pages. I'll get back with an update when I get home. cheers Ian Chivers
From: Jason Blevins on 12 Aug 2010 09:50 On 2010-08-12, Ian <ian_d_chivers(a)yahoo.co.uk> wrote: > On Aug 11, 3:50 pm, Jason Blevins <jrble...(a)sdf.org> wrote: >> The Fortran Forum's status table for Fortran 2003 support for Intel >> seems to be somewhat dated. According to Intel's article on Fortran >> Language Standards [1] from June 2009, the following are supported in >> version 11.1 but are still listed as not supported in the table: >> >> * Deferred bindings and abstract types >> * Type-bound procedures >> * ABSTRACT attribute >> * DEFERRED attribute >> >> I've verified this by compiling some simple example programs using >> these features. So, it seems to me that the following rows in the >> Intel column should have a "Y" instead of an "N": >> >> * Procedures bound by name to a type >> * Deferred bindings and abstract types >> >> I'm happy to contact Ian (if he's not reading this already) but I >> wanted to get some independent verification first, just in case I'm >> misunderstanding something. Perhaps someone knows of other rows that >> should be updated as well? >> >> [1]:http://sdf.org/l/c43 > > Thanks for bringing this up. > >> * Procedures bound by name to a type > > should be partial. > > I'll correct in the next edition. > >> * Deferred bindings and abstract types > > The information I use for the Intel entry > is provided by one of the compiler team > at Intel, Stan Whitlock. I should have changed > no to partial in the entry above, but missed it > in the last update. > > > The other is still no as far as I can determine. > I'm on holiday at the moment and don't have complete > access to my records. The abstract type, deferred binding example on page 284 of MR&C compiles without issue for me, so perhaps there is at least partial support now? I didn't go as far as to write a full test program though, so I can't verify that it works as expected. > Most of the information in the table comes from compiler people > rather than their web pages. > > I'll get back with an update when I get home. I appreciate your efforts in maintaining these very useful tables! Not having any inside information, it's hard to distinguish P from Y on my end (since a single working example doesn't imply full support, nor does a list of features on the website), so thanks for taking a closer look at these. -- Jason R. Blevins Assistant Professor of Economics Ohio State University http://jblevins.org/
|
Pages: 1 Prev: F2008 implied-shape array and scalar initialization Next: Asking for a good book |