[Roldan is our benchmarking expert. He post-doced under Dongarra and has worked
on LAPACK++ and Sparse BLAS.]
Roldan Pozo wrote:
>
> The Linpack benchmark is the most popular program for linear
> algebra. If you want to compare Fortran instrinsics to the BLAS,
> the easiest thing is to write a simple program to time these.
> (The program should take the matrix size (M,N) is inputs, generate
> an array of random values, and repeat the operation enough
> times to remove clock granularity effects.)
>
> --Roldan
> > Van Snyder wrote:
> > >
> > > What are your favorite programs to measure the performance of linear
> > > algebra procedures, especially vector and matrix inner product
> > > procedures? I'm interested to compare Fortran intrinsic DOT_PRODUCT
> > > and MATMUL to xDOT and xGEMM, respectively.
> > >
> > > Best regards,
> > > Van Snyder
--
William F. Mitchell
Mathematical and Computational Sciences Division
National Institute of Standards and Technology
[log in to unmask] http://math.nist.gov/~mitchell
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|