Harald J. Ehold

  • Citations Per Year
Learn More
Numerical algorithms written in HPF often have to pay performance penalties, when they are compared with optimized numerical libraries on single processors. Obviously, this is a problem for achieving high parallel performance, too. In order to overcome these e ciency problems a method to utilize existing numerical single processor software for(More)
Portable and eecient ways for calling numerical high performance software libraries from HPF programs are investigated. The methods suggested utilize HPF's EXTRINSIC mechanism and are independent of implementation details of HPF compilers. Two prototypical examples are used to illustrate these techniques. Highly optimized Blas routines are utilized for(More)
HPF was originally created to simplify high-level programming of parallel computers. The inventors of HPF strove for an easy-to-use language which was intended to enable portability and e ciency. However, up until now the desired e ciency has not been reached. On the contrary, HPF programs are notorious for their poor performance. This paper provides a(More)
  • 1