From: Alan W. I. <ir...@be...> - 2017-03-22 21:36:48
|
On 2017-03-22 07:55-0000 Arjen Markus wrote: >> -----Original Message----- >> From: Alan W. Irwin [mailto:ir...@be...] >> Sent: Tuesday, March 21, 2017 8:00 PM [...] >> Our experience is the PLplot Fortran binding and examples work fine with the NAG >> fortran compiler (nagfor), Intel Fortran compiler (ifort) and the Fortran compiler >> (gfortran) that is part of the gcc suite of compilers. But has anybody here tried other >> well-known Fortan compilers such as those from Absoft or the Portland Group? My >> understanding is CMake supports both the Absoft and Portland Group Fortran >> compilers so the question really boils down to whether those compilers have the >> necessary support for Fortran 2003. >> >>> AM: I am pretty sure they do. The standards' support for a number of compilers is regularly compared by Ian Chivers and Jane Sleightholme for the ACM Fortran Forum and the 11 compilers they list all have full support for the F2003 C-Fortran interoperability [....] Hi Arjen: I followed up by looking carefully at the part of the table in <http://www.fortranplus.co.uk/app/download/23704631/fortran_2003_2008_compiler_support.pdf> corresponding to C interoperability for the 11 Fortran compilers that are covered in the report (i.e., Absoft, Cray, g95, gfortran, HP, IBM, Intel, NAG, Oracle, PGI, and Path scale). Those first 10 have a "Y" (yes) for all C interoperability features while Path scale has a "Y" for all features other than "Interoperability of global data" where it has a "N" (no). So from your perfect results (no build warnings or errors, no run-time errors, and perfect PostScript difference results) for a compiler (NAG) that is notorious for being standards compliant, my and your good results for gfortran and ifort (Intel), and from the results from this report it appears there is a good chance that Absoft, Cray, g95, HP, IBM, Oracle, and PGI will also produce good results for our new Fortran binding for PLplot. And the Path scale Fortran compiler might also produce good results depending on whether or not our Fortran binding uses the "Interoperability of global data" feature. However, it would be nice to get these anticipated good results confirmed for the eight compilers we have no results for yet, and also it would be good to get confirmation of our own good results for gfortran, Intel, and NAG. So I urge anyone here with access to any of those 11 compilers to give them a try with PLplot-5.12.0 and let us know the results. Also, getting back to the original Silverfrost Fortran topic, I think it is telling that that vendor hasn't bothered to register with the authors of the above well-known report despite a general invitation (at the start of that report) for any Fortran compiler vendor to participate. Alan __________________________ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __________________________ Linux-powered Science __________________________ |