On Wed, 2002-11-13 at 16:39, Rui Miguel Seabra wrote: > On Wed, 2002-11-13 at 15:23, Margaret_Doll wrote: > > I just upgraded my systems to RedHat 8.0 including the kernel patch on > > their web page. > > The systems are at 2.4.18-17.8.0. > > I get the following error when I run PG 4.0 on my systems. > > /usr/bin/ld: cannot open > > /usr/lib/gcc-lib/i386-redhat-linux/2.96/crtbegin.o: > > No cush file or directory > > My my, I guess the answer is right here: 2.96. It will only work with > RedHat 7.x which uses gcc 2.96, or other likewise distributions. > > > I understand that we have a campus license for PGI workstation or Pgi > > Server. > > How do I get my systems running the pgi compilers? > > Downgrade and suffer the consequences of proprietary software or get rid > of PGI: RedHat 8.0 gives you the magnific GNU Compiler Collection 3.2 > http://gcc.gnu.org/. > > >From the web page: > > GCC is the GNU Compiler Collection, which currently contains front ends > for C, C++, Objective C, Fortran, Java, and Ada, as well as libraries > for these languages (libstdc++, libgcj,...). > Please, trying to force GCC on people's throats when it simply doesn't meet the requirements does NOT help free software. For instance Gcc's code does not even come close to code generated by Intel's ICC. Partly because ICC "cheats" it inlines functions at -O2 level while GCC does it at -03 level. But ICC knows how to do interprocedural optimizations and Gcc not. That makes for icc code often being 50 to 100% faster than GCC even when using gcc best combo of. Another reason to use PGI is Fortran 90. Gcc only understands Fortran 77. And Ada in Gcc 3.2 is so broken that I have had to downgrad to the gnat who was based on gcc 2.8 > I recommend getting rid of PGI. > I recommend you stop giving such advices: when free software is not as good as a proprietary alternative but still good enough you can recommend it using its other advantages as arguments. When it does not meet minimum requirements (eg Fortran 90 or code speed when it is crucial (for instance when sales depend on performance)) you can get only two things: people try your solution, get burned and then free software gets a bad name, or someone with enough knowledge points the drawbacks of the free solution and next time you try to push a free solution who also happens to be the best you will be ignored because people will regard as a fanatic JFM -- Psyche-list mailing list Psyche-list@redhat.com https://listman.redhat.com/mailman/listinfo/psyche-list
