[comp.misc] price/performance

mct@praxis.co.uk (Martyn Thomas) (11/16/90)

Every five years, your dollar buys about 10-times as many MIPS, if you look
at factory-gate microprocessor prices.

Has anyone got historic ratios for the improvement in price/performance for
IBM or any other large processor manufacturer?

I need them for a lecture on Nov 21st.

Thanks
-- 
Martyn Thomas, Praxis plc, 20 Manvers Street, Bath BA1 1PX UK.
Tel:	+44-225-444700.   Email:   mct@praxis.co.uk

herrickd@iccgcc.decnet.ab.com (11/20/90)

In article <5517@newton.praxis.co.uk>, mct@praxis.co.uk (Martyn Thomas) writes:
> Every five years, your dollar buys about 10-times as many MIPS, if you look
> at factory-gate microprocessor prices.
> 
> Has anyone got historic ratios for the improvement in price/performance for
> IBM or any other large processor manufacturer?
> 
> I need them for a lecture on Nov 21st.
> 
> Thanks
> -- 
> Martyn Thomas, Praxis plc, 20 Manvers Street, Bath BA1 1PX UK.
> Tel:	+44-225-444700.   Email:   mct@praxis.co.uk

Back before 1976 Herb Grosch was a columnist in Computerworld, a weekly
news rag aimed at MIS people.  He would occasionally write about
Grosch's Law, which was a formula for such things.  The formula
had quite good predictive value for a long time.  Methinks I remember
reading someone's comments on the current applicability of Grosch's
Law in an IEEE or ACM publication within the last two or three years.

dan herrick
herrickd@astro.pc.ab.com