Path: utzoo!attcan!uunet!auspex!guy
From: guy@auspex.UUCP (Guy Harris)
Newsgroups: comp.arch
Subject: Re: The NeXT Problem
Message-ID: <22@auspex.UUCP>
Date: 19 Oct 88 16:26:04 GMT
References: <26435@ucbvax.BERKELEY.EDU> <5498@juniper.uucp> <250@dataspan.UUCP>
Reply-To: guy@auspex.UUCP (Guy Harris)
Organization: Auspex Systems, Santa Clara
Lines: 44

>I am getting sick and tired of this RISC/CISC battle. Come on guys, measure
>RISC instructions against CISC microcode level instructions and they work out
>to about the same thing.

No, measure how fast a given application runs; that's what counts.

Note also that I, or a compiler, can get at the RISC instructions,
while neither I nor a compiler can generally get at a CISC machine's
microinstructions (assuming it has microinstructions).  Also note that the
microinstruction sets for two different implementations of a given CISC
architecture are often (usually?) not the same, which means if you
*could* get at the microinstructions you would have to recompile (or
rewrite) microprograms to get them to run on the new implementation.

>As far as I can see, it appears that we are always battling memory
>speeds.

Not clear.

>Not trying to start a flame war, but 030's are faster than Sun 4's.

To which '030 machine, and to which Sun-4, are you referring?  At the
time the Sun-4/260 came out, no available '030 machine was faster
because there weren't any '030 machines.  What happens when the next
Sun-4 comes out?  Note Allen Baum's argument that RISC lets you get to
market faster with a machine of a given speed, which means that by the
time a CISC machine catches up you're on your way to a new faster
implementation.

Also, you might compare '030s against MIPS-based machines; are they
faster than them, as well?

>I puke trying to write assembly on RISC machines.

Fortunately, I rarely had to do so, and these days fewer and fewer
people writing applications have to do so.

These days, "ease of writing assembly code" is less and less of a
figure of merit for an instruction set.

>I love CISC though, because you find your bottleneck, and just jump in
>and start ripping some things out, and rewrite the routine..

You can do that with RISC as well.