Fortran Programming Language :: Intel Fortran compiler version 10 released


	Intel Fortran compiler version 10 released Version 10 of the Intel compiler for Windows and Linux has been released. The Windows version is now stand-alone. It is a "Self- contained product includ[ing] Microsoft Visual Studio 2005 Premier Partner Edition" http://www.intel.com/cd/software/products/asmo-na/eng/compilers/27883... . There a few Fortran 2003 features, such as C interoperability, but it is far from a full F2003 compiler. I don't work for Intel. Beliavsky wrote: > There a few Fortran 2003 features, such as C interoperability, but > it is far from a full F2003 compiler. I don't work for Intel. Does anyone have a fair idea about which of the various Fortran compilers around are now "closest" (however that is to be understood) to having implemented the full F2003 standard? Just curious ... :-) -- -+-Ben-+- On Jun 7, 2:43 pm, Ben Hetland <ben.a.hetl@sintef.no> wrote: > Beliavsky wrote: > > There a few Fortran 2003 features, such as C interoperability, but > > it is far from a full F2003 compiler. I don't work for Intel. > Does anyone have a fair idea about which of the various Fortran > compilers around are now "closest" (however that is to be understood) to > having implemented the full F2003 standard? I'd guess NAG. Quoting the June 6, 2007 issue of NAGNews (a free email newsletter from NAG), "A beta version of NAGWare f95 Release 5.1 has been used in the preparation of a new book. Author Stephen J. Chapman says "NAG is way ahead of other PC Fortran vendors in implementing the object-oriented features of Fortran 2003 and its tool has been invaluable in preparing the new materials in this book." Fortran 95/2003 for Scientists and Engineers, 3rd ed. is published by McGraw-Hill, ISBN 978-0-07-319157-7. Beliavsky <beliav@aol.com> wrote: > On Jun 7, 2:43 pm, Ben Hetland <ben.a.hetl@sintef.no> wrote: > > Beliavsky wrote: > > > There a few Fortran 2003 features, such as C interoperability, but > > > it is far from a full F2003 compiler. I don't work for Intel. > > Does anyone have a fair idea about which of the various Fortran > > compilers around are now "closest" (however that is to be understood) to > > having implemented the full F2003 standard? > I'd guess NAG. Quoting the June 6, 2007 issue of NAGNews (a free email > newsletter from NAG), NAG was to my knowledge the first out of the gate with several of the basic features, but it seems to me that they have stalled quite a bit for several years. Their version 5.0 seemed an impressive start. But then I kept waiting for more. I think I recall it was something like a year and a half or so between 5.0 and 5.1 and I was really dissappointed at the small amount of progress; some, but it was sure small compared to the major stuff in 5.0. It is pure speculation, but I note the correlation between the slowdown and Malcolm being over in Japan for a few years. (He's also been editor of the standard the last few years after I stopped, and I know that's a big job). I'm suspicious that he has had to spend more of his attention on other things... and Malcolm's darned good - hard to adequately substitute for. Ths single thing that I miss most in NAG is allocatable scalars. That is a pretty big limitation on the object oriented stuff, as anything polymorphic has to be either allocatable or pointer. You can, of course, always use pointers, but that has a host of the usual problems, which pretty much destroy many of the advantages of the object oriented stuff in my view. Polymorphic things mostly "want" to be allocatable. And, of course, you can make everything arrays, but that's not exactly viable for real work or anything other than demonstrating basic principles, in my view. Plus, with the allocatable scalars, you get variable-length strings done "right", or any way a whole lot better. I haven't tried it, not having a Cray handy here at home :-), but the feature table in the latest Fortran Forum makes Cray look pretty much in front at the moment. -- Richard Maine \| Good judgement comes from experience; email: last name at domain . net \| experience comes from bad judgement. domain: summertriangle \| -- Mark Twain