There is a nice little page on the Digital site:

http://www.digital.com/alphaserver/performance/vupbs_297.html

Where it states:

VUPS ... was the Geometric Mean of the ratios of the execution times
of 99 specific CPU-intensive benchmarks, including 71 Fortran benchmarks,
4 COBOL benchmarks and 24 Lisp benchmarks.

Digital generally published the overall VUP ratings for systems, but
did not usually publish details on each of the individual benchmarks.

VUPS could not be used to compare hardware from different vendors, 
because the suite was not made available to different vendors.

VUP ratings were CPU-centric, and did not provide very much information
to users who were primarily concerned about disk I/O, graphics, etc.

Interestingly, Digital discontinued the VUP ratings with the Alpha.
See that page for more info.

Doug



On Wed, 2 Feb 2000, emanuel stiebler wrote:

> > On Wed, 2 Feb 2000, emanuel stiebler wrote:
> >
> > > Anybody here know (or remember ;-), how this was measured, benchmarked ?
> > >
> > > I know, that the 11/780 should be 1 VUP, but how they benchmarked the
> > > newer machines ?
> >
> > I'm not really answering your question here, but I'll tell you what I've
> > done.  I tend to just take the Dhrystone benchmark and run it on a known
> > VUP VAX.  I take the resulting dhry/s value, divide it by the known VUP
> > value for the system, and then use the result as a dhry/s to VUP ratio.
> >
> > >From here you can benchmark an unknown system using the Dhrystone
> > benchmark can convert to VUPs with the ratio.  I've found that the
> > calculated VUP results match up pretty well with the values advertised
> > by DEC.
> 
> What you're telling (writing ? ;-) is, that VUPs has nothing to do with
> floating point performance ? No I/O ?
> 
> cheers,
> emanuel
> 
>