[Simh] VUPs?

[Tim Shoppa]

Dave writes:
> The number generated by VUPS.COM does NOT necessarily correlate to the
> Digital's official published VUP rating for VAXen or Alphas. I believe
> SRI has a disclaimer to that effect.

I don't think DEC ever officially released their VUP benchmark suite to the wide world.

Digital Review, back in the days, had their own VUP test suite that gave similar
but not identical numbers. Many put much more weight on the Digital Review
benchmarks because they were published.

Some tests are more I/O bound, others more compute bound, some more integer
bound, some more floating-point bound, and these did not all scale identically
between "hardware" VAXen and they do not scale all identically between simulated
VAXen either.

Interestingly enough, my understanding is that the platinum-iridium VAX VUP, the
VAX config that is by definition 1.000000 VUPS, is a 11/780 with a specified set
of Massbus disks etc running a specific old version of VMS and layered products.
Same hardware with different hardware versions, or even a different SYSGEN,
gives slightly differnt numbers.

Don't know the exact versions, seems a little academic unless someone wants to
actually build a config which is a platinum-iridium VAX VUP and put it in a vault next
to the platinum-iridium kilogram in Sevres, France :-).

Tim.