At 12:19 PM 2/2/00 -0700, emanuel wrote:
>Is there enough interest here to collect dhrystones of the different
>machines ?

There were some numbers somewhere on the net. However Dhrystones were very
susceptible to compiler optimizations so ideally you should create an
executable and then re-use that executable on every machine (rather than
recompile it). 

That being said, SPECMarks work better because they floating point as well
as integer performance and there are numbers for other systems as well.

On the third hand, the only reason one _might_ care about documenting how
slow a VAX is, is to compare with still slower machines like the PC-XT. :-)
The VUPs are documented and available in the literature so re-creating them
appears to me at least to have little value.

--Chuck