[Simh] Questions on SIMH VAX
Robert G. Schaffrath
robert at schaffrath.net
Thu Jun 10 21:36:56 EDT 2010
One of the problems with calculating Instructions Per Second on the VAX
and using that figure as a benchmark was the fact that some of the VAX
instructions were very complicated and could take upwards of eight clock
cycles to complete. On RISC systems, where the goal was having all
instructions execute within a defined number of clock cycles,
calculating MIPS was easier and perhaps more relevant. When Digital
reworked the VAX architecture in the mid-1980's and moved some of the
more complicated and less frequently used instructions from microcode to
software (as well as removing native PDP-11 instructions), the MIPS
definition may have become a better benchmark. Still, for customers who
had a VAX 11/780 or were accustomed to its performance, the VUP at least
gave you a concept as to how much faster your bright new shiny VAX 6xxx
was compared to the original 11/780.
Robert
On 6/10/2010 9:21 PM, Rich Alderson wrote:
>> Date: Thu, 10 Jun 2010 11:45:35 -0700
>> From: "Zane H. Healy"<healyzh at aracnet.com>
>>
>
>> I thought a VAX-11//780 was both 1 MIP, and 1 VUP.
>>
> Zane,
>
> ITYM "MIPS". It's not a plural, it's an abbreviation of "second".
>
> Anyway, although Digital's marketeers tried to claim that the 780 ran at
> 1MIPS, it was clearly doing about half that in any decent benchmark. That's
> the reason they came up with the VUP, very quickly. AFAIK, no one else ever
> bothered to use the VUP as a measure, even if they were into MIPS.
>
> Rich
> _______________________________________________
> Simh mailing list
> Simh at trailing-edge.com
> http://mailman.trailing-edge.com/mailman/listinfo/simh
>
More information about the Simh
mailing list