README.optimize_for_cpu

8/2021       Optimize for a specific CPU

The wgrib2's makefile assumes a generic CPU.  For 64-bit AMD or Intel 
cpus on a 64-bit OS, that means wgrib2 will be compiled for the 64-bit 
instruction set first introduced by AMD in 1999.  That means the latest 
instructions like AVX+ will not be used.  The advantage of using the older
instruction set is that the executable doesn't have to be recompiled
to a compatible OS. 

In my experience, optimizing wgrib2 to new instructions provided limited
gains.  Speeding up wgrib2 by recompiling is limited because

  (1) Wgrib2 is often limited by I/O speeds.
  (2) Wgrib2 has limited parallelization in packing and unpacking.
  (3) Wgrib2 uses OpenMP to parallize the loops, and this limits the 
      parallelization that can be achived by modern SIMD instructions.
  (4) Wgrib2 limits itself to OpenMP 3.1.  Later versions of OpenMP
      can enable SIMD instructions and run on GPUs.  However, portability
      is an issue.

If you want to try optimizing wgrib2, add the following before compiling

export CPPFLAGS='compiler-directives' (previous version of README had a typo, CFFLAGS=..)
export FFLAGS='compiler-directives'
export CC=C-compiler
export FC=Fortran-compiler
make


For compiler directives, the -march=(CPU) is easy to use.