Table i presents a comparison of our radix sort with that of an implementation using only single phase routing by Alexandrov et al. which we will refer to as the AIS code. Performance of the latter code, which had been optimized for the Meiko CS-2, is given in [2]. Note that the AIS implementation is based upon the original version by Dusseau ([20,18]). Also, all codes in this comparison have been written in the Split-C language [17]. Our algorithm is referred to as BHJ.
Table i: Total Execution Time for Radix Sort on 32-bit Integers (in seconds), Comparing the AIS and Our Implementations