[en] The NAMMU program has been converted from an IBM double precision version to a CY205 single precision version. Due to different dialects among compilers, differences on system levels and machine dependencies some changes to the code have been made. They are documented in this report. A scalar optimized version is delivered. Direct vectorization did not improve performance. Changes in program logic is necessary to take advantage of the the vector capabilities. That and a study of the input-output operations are needed to improve performance. (author)