Design and implementation of matrix algorithms of systolic array processor

Design and implementation of two impor- tant matrix operations namely matrix multi plication and matrix Inversion using geo metric arithmetic parallel processor (GAPP) array is described. These matrix operations are implemented oh an IBM PC/XT compatible with an add on card con sisting of two GAPP clips. Together, these two clips have an array of 12 X 12 processing elements. The implementation is done using GAL - GAPP algorithm lang uage. Performance evaluation shows that this idea can be extended to large arrays.