LINPACK
LINPACK initially designated a numerical library for solving systems of linear equations. Later, the name has been used for a program for measuring the speed of a computer. In the original version of this measurement program almost all the work is done in two subroutines called from the program library.
The program library LINPACK was developed in the 1970s and 1980s by Jack Dongarra, Cleve Moler, Gilbert W. Stewart and others, and was later replaced largely by LAPACK.
The measurement program LINPACK is used for measuring the performance of supercomputers since the solving of linear systems plays a crucial role in there often implemented implicit solution methods. The result is given in floating point operations per second ( FLOPS ) and published in the TOP500 list of fastest supercomputers. LINPACK programs are usually written in Fortran, but there are also versions in the C, C , Pascal, Java, and other programming languages.
Criticism
Lately, the comparison with LINPACK often gets into criticism because certain architectures perform better with him, and he therefore does not provide comparable results. Especially systems that are equipped with relatively low memory bandwidth and capacity, have it too easy, thanks to its pure CPU power to get to the top of the list, even though they used only for very specific problems, but not for large, realistic simulations can be.
As a current example of this is the super computer MDGRAPE - 3 ', which performs in the Japanese RIKEN Research Center for Molecular Dynamics calculations. Its computing power is more than a petaflop, thus the tenth- placed hosts the Top 500 beats IBM's Roadrunner. Nevertheless, it can not be included in the TOP500 list, as the LINPACK benchmark can not run on it.