Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!columbia!rutgers!andromeda!argus!ken
From: ken@argus.UUCP (Kenneth Ng)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Mainframe vs Micro
Message-ID: <722@argus.UUCP>
Date: Thu, 22-Jan-87 01:43:19 EST
Article-I.D.: argus.722
Posted: Thu Jan 22 01:43:19 1987
Date-Received: Thu, 22-Jan-87 06:22:19 EST
References: <658@imsvax.UUCP> <1490@cit-vax.Caltech.Edu>
Organization: NJ Inst of Tech., Newark NJ
Lines: 46

In article <1490@cit-vax.Caltech.Edu>, li@tybalt.caltech.edu (James C. Li) writes:
> Would someone DEFINE what a Micro is and what a Mainframe is?  This argument
> seems to be going in circles because as Micro's increase in power, so do 
> Mainframes.  It isn't fair to compare Micro's of todays to Mainframes of the
> 60's.  Also, what is a Mini?  And what about a Supercomputer?  What about the
> Hypercube(claim is 100 68020's in parallel can out compute a Cray 1)?
> (Yes, the right size/type for the right application, but what is similar about
> these systems that CAN be compared, or can't they?)
> li@tybalt.caltech.edu (James Li)           __   __
> (also li@citromeo.caltech.bitnet)            | |
> "Official KANK symbol"         ------>       | |

We could use the old definition: a micro you can lift with two hands,
a mainframe you can't.

The Cray supercomputer line has supposedly been outpaced by several computers
in recent years, especially by parallel and vector processing machines. But
one must realize that no matter how much one parrellels stuff, there will
always be stuff done in serial fashion.  And that is where the Cray really
excells, with one of the fastest (is it the fastest?) scaler processors around.

By the way, here is an actual benchmark done here at NJIT on various computers.
The benchmark is a compute bound matrix solution to find the capicatence of
a square plate via point charge analysis.  The speed killer is to solve a
4x4, 9x9, 16x16, up to a 400x400 floating point matrix.  An IBM PC (without
8087) was stopped after a week, it didn't complete.  An AT&T 3b5 with an
average load of 5 was stopped after 3 days (system crash).  An IBM 4361
with an average load of 3 users took about 9 hours.  A Univac 90/80-4
with about 80 users took 10 minutes.  Need I say more about the power
of mainframes?  And the Univac is 1970's technology to boot.
Anyone have a Cray I can borrow to find out how long it will take there?  

Now, there are some differences that must be taken into account.  First,
was the quality of the compilers.  The Univac compiler was NICE and
generated VERY tight code.  The 4361 compiler I have strong reservations
as to how good it is.  Perhaps one year when I get bored I'll code the
routines all in assembly and then run the tests.

-- 
Kenneth Ng: Post office: NJIT - CCCC, Newark New Jersey  07102
uucp !ihnp4!allegra!bellcore!argus!ken
     ***   WARNING:  NOT ken@bellcore.uucp ***
bitnet(prefered) ken@orion.bitnet

Gillian: "Are you sure you won't change your mind?"
Spock: "Is there something wrong with the one I have?"