hkbirke@mit-amt (Hal Birkeland) (10/20/88)
NOTE: this article was originally posted in alt.next... I have replied here since alt.next should be disappearing soon... In article <6119@quacky.mips.COM>, dennis@mips.COM (Dennis Franklin) writes: > In article <5498@juniper.uucp> chari@juniper.UUCP (Christopher Michael Whatley) writes: > >... Doesn't 20 RISC MIPS equal about 5 CISC MIPS? > > > The "MIPS" figure that is published by RISC chip/computer manufacturers > IS what you term "CISC MIPS". It is usually stated as "VAX MIPS". I.e., > 1 VAX 11/750 = 1 MIP. well, almost (actually a real good try). But according to DEC's pamphlets dating from the 11/7xx series machines, the 11/750 was not rated at 1.0 MIPS but at .7 MIPS. The 11/780 was rated at 1.0 VAX MIPS. For comparison, a uVax II is a 0.7 MIPS machine. Just remember that processor speed isn't everything, i/o bandwith is also important (not that NeXT's have problems here). --hal {backbone}!mit-amt!hkbirke hkbirke@media-lab.media.mit.edu
hedrick@athos.rutgers.edu (Charles Hedrick) (10/20/88)
>1 VAX 11/750 = 1 MIP.
?? As far as I know, the Standard MIP is a 780. However it's probably
better to ask people to say exactly what they are comparing with, i.e.
what machine, OS, and compiler. Recently I've started to see "Ultrix
MIP's" quoted. Apparently this is a comparison with a VAX 780 running
Ultrix. The theory is that it's only fair that a Unix system should
compare against DEC's Unix. However if this is an excuse to use the
old f77 instead of DEC Fortran, the result is being seriously biased.
And even then, there's going to be a range, depending upon the
application. E.g. Sun rated their 4/280 at 10 VAX MIP's. I think
most people now believe that 8 is more like it, but more like 5 for
Fortran, and probably considerably more than 10 for Lisp.