*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Since gfortran has been subject to some criticism in this forum, I tried
compiling the new sources of shelxc/d/e currently on my server with it.
Unlike most (all ?) ccp4 program and the older shelx codes (e.g. shelxl)
they cannot be compiled with g77 because they use a few Fortran-90
facilities (in particular allocatable arrays). I used the gfortran that
came with SuSE 10.0 and chanced my luck with:
gfortran shelxc.f -O5 -ffast-math -mtune=pentium4m -static -o shelxc
and similarly for shelxd and shelxe. The static executables produced in
this way were tested on three random different computers (3.06 GHz P4,
1.2 GHz AMD and 2.15 GHz Pentium M laptop) with Linux versions 2.6.11,
2.2.16 and 2.6.13 resp. All tests ran correctly allowing for the effects
of the random number seeds. The run times varied appreciably, but the
standard Intel ifort executables were a little faster on average on the
genuine Intel CPUs and gfortran was a little faster on the AMD; the
laptop easily outperformed the two workstations in all tests.
That looks to me to be already a very respectable result for gfortran!
George
--
Prof. George M. Sheldrick FRS
Dept. Structural Chemistry,
University of Goettingen,
Tammannstr. 4,
D37077 Goettingen, Germany
Tel. +49-551-39-3021 or -3068
Fax. +49-551-39-2582
