something to read
The best benchmark remains your own application. I assume you have profiled it and know
where it spends its time, and have optimised it where possible; running it will at least ensure
that the machine, operating system, and compiler exist, a factor often overlooked amidst
a profusion of numbers and promises.
Benchmarking Computers for HEP
CERN, Geneva, Switzerland
PPM benchmark results (CFD application)
quite frequently I have come accross the following book title, and
never know the complete set of author, title, who published it ....
the title is: 'Numerical Recipes'
Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.
Numerical recipes in FORTRAN: the art of scientific computing
Second Edition, 1992
Cambridge University Press, Cambridge
those were the days...
We were slightly worried when we discover the OS was written in what
appeared to be a dialect of Fortran II. Internal evidence indicated
that much of the OS was originally written by physicists at LLL;
which possibly explains the use of Fortran :-)
Geoff Lane's experience with Cyber205
note on efficient algorithms
We suspect it would take considerably longer than the age of the
universe to be able to compute the latter result using the complete
weight space method.
Wybourne, B.G., 1994, Comput. Phys. Commun., 83, 332