len@elxsi.UUCP (Len Mills) (06/02/87)
Results on the Dhrystones benchmarks are heavily influenced by the efficiency of the various string handling functions. Much of the CPU-bound processing consists of string manipulations. The results being reported for benchmarking times may have more to do with the string functions than with the compiler efficiency. Anyone for assembly versions of those routines? I don't even have THE BOOK, yet.