Re: performance / f2c bug

Harald Koenig (koenig@tat.physik.uni-tuebingen.de)
Fri, 14 Jun 1996 13:50:16 +0200 (MET DST)


> > My question: has anyone ever used the RedHat 3.0.3 (or 2.1) linux
> > f2c on the Alpha successfully (for non-trivial programs)? Has anyone
> > an idea what is wrong - you obviously see what makes f2c fail in
> > the code.
>
> Tried, without success. Then I gave them a way to
> compile binaries at DEC OSF/1, and run them at Linux...

I've compiled f2c myself (using "-mieee"; didn't try the binary in 3.0.3)
but used the standard libc and couldn't find any problems then compiling
a few "mid sized" fortran progs (several 1000 lines) from my old days ;)

for a "real" program I usually use (for benchmarks only these days) SHELXL-93
(CRYSTAL STRUCTURE REFINEMENT, ~20k lines fortran giving ~42k lines C)

f77 1030.31user 2.09system 17:14.05elapsed 99%CPU
f90 1113.64user 2.24system 18:37.76elapsed 99%CPU
cc 1318.54user 6.93system 22:07.94elapsed 99%CPU
gcc 1548.95user 8.06system 25:57.20elapsed 99%CPU

where "f77", "f90" "cc" are DEC's native compilers (just using -O2 or similar;
no special options; using f2c for "cc" of course; binaries have been linked
statically for Linux; but the shared libs version runns too if you have their libs
available on Linux). "gcc" is gcc-2.7.1 on Linux.

all times are on a XL266 (21064A-266) with 2MB cache and 48MB ram
(the main dataset won't fit into 2MB cache) and almost all operations
are done on REAL*4 (float) to save memory on good old FORTRAN's machines...

Harald

-- 
All SCSI disks will from now on                     ___       _____
be required to send an email notice                0--,|    /OOOOOOO\
24 hours prior to complete hardware failure!      <_/  /  /OOOOOOOOOOO\
                                                    \  \/OOOOOOOOOOOOOOO\
                                                      \ OOOOOOOOOOOOOOOOO|//
Harald Koenig,                                         \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik                              //  /     \\  \
koenig@tat.physik.uni-tuebingen.de                     ^^^^^       ^^^^^