For a general tutorial on Open MP, check out www.openmp.org


  Re.: how to compile and run your code on the picard machine, 
here's how you do it for FORTRAN. I use fortran90, but it would work for fortran77, 
as well (you compile with 'f77' instead of 'f90' in that case). 
For the sequential execution, compile program matrix.f, you do the following:


f90 -O3 matrix.f


  For the parallel execution using Open MP, in addition to adding an appropriate 
directive in the code (smt. like "c$omp parallel ..." - see lecture notes for details), 
you compile with


f90 -openmp -O3 matrix.f


  The purpose of "-O3" is to instruct the compiler to aggressively optimize the code, 
for the fastest execution.


  To run the compiled code and measure the execution time, type


time a.out


  It works similarly for C, except that the compiler directives need to have an 'x' in 
front, as in:


cc -xopenmp -xO3 filename.cs