Asri-unix.1238
net.works
utzoo!decvax!cca!sri-unix!Hank
Fri Apr 16 06:38:54 1982
bogus benchmarks
The most flagrant violator of benchmark reasonableness that I've seen lately
is Perkin-Elmer.  They had an add comparing their machine to a VAX-11/780.
The machine supposedly executed one of the benchmarks >1000 times faster than
the VAX.  How was this possible.  Well the benchmark didn't generate any
output, so the compiler optimized it to a NOOP, which could be executed
1000 times faster...