rick@pcrat.UUCP (05/20/87)
I am suspicious of the Microport UNIX results I just sent out in the Dhrystone posting. I'm hoping someone can recheck these results. The reason I'm suspicious is that the 6300 PLUS UNIX results* are around 1200 and some of the Microport results are claiming around 2100. Clock differences aside, there is clearly something wrong here, since both operating systems have the same compiler. As a sanity check, please compile the benchmark using time() to measure time, as well as times() (TIME and TIMES #defines); sometimes the HZ parameter is not what the operating documentation claims. Microport claims 60 HZ, but I've heard rumors that the time of day clock loses time on that operating system (potentially meaning that the HZ parameter IS NOT precisely 60 HZ). It may turn out that there's no effective way to measure performance on Microport, except with a stopwatch. * I know the PLUS results are correct, I did them myself. -- Rick Richardson, President, PC Research, Inc. (201) 922-1134 (voice, nights) OR (201) 834-1378 (voice, days) seismo!uunet!pcrat!rick ihnp4!castor!pcrat!rick