🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: net.arch
Article #3094

Let's start a trend! (Performance measurement -- units)

#3094
From: Eugene Miya N.
Date: Tue, 04 Nov 1986 01:27
48 lines
2588 bytes
%A C. Bratten
%A R. Clark
%A P. Dorn
%A R. Grant
%T IBM 3090 Engineering/Scientific Performance
%I Washington Systems Center
%C Gaithersburg, MD
%R Technical Bulletin GG66-0245-00
%D June 1986

Frank McMahon (LLNL developer of the Livermore loops) and I were discussing
performance when he pulled this new paper out.  It seems to imply that
IBM developed vector processing just as IBM "developed" virtual memory,
caches, and several other things.  One section of the paper notes how
IBM determined that 64 elements in a vector register were the optimum
number to have (gee, I wonder if Seymour Cray really has a blue suit
hidden in a closet...;-).

Anyway, I would like to see if we can start a fad.  We had pet rocks,
we had :-), we had little yellow signs, and we had the vaporware fad.
I suggest we adopt the following prefix unit: "macho" in place of
"peak Mega."  This differs from sustained Mega.  For example: we have
MFLOPS: this is not a Mega, this is a Macho FLOPS rating.  Bits move at
Macho Bytes per second, or you might be a Macho IPS person.  Don't
express that "peak kilobytes transfer rate" that's 0.XXX Macho bytes.
Gigabytes of storage? (Apologies to Chris Lloyd's "Back to the Future"
character who said it nicely), say "thousands of Macho-bytes."

I'm willing to shelve the old "Sagans" (billions and billions) as a unit
if we can really get people thinking about performance as hype.

Say it to yourself:
	Macho Bytes per second
	Macho FLOPS
	Macho IPS
	1220 Macho volts
	0.000042 Macho solutions

I would like to thank to unknown individual who first thought about this
prefix and who passed it to Bob Brown at RIACS two years or so ago.

From the Rock of Ages Home for Retired Hackers:

--eugene miya
  NASA Ames Research Center
  eugene@ames-aurora.ARPA
  "You trust the `reply' command with all those different mailers out there?"
  {hplabs,hao,nike,ihnp4,decwrl,allegra,tektronix,menlo70}!ames!aurora!eugene
Technical Details
Message-ID: <722@nike.UUCP>
Path: arc1-ca.pugleaf.net!utzoo!mnetor!seismo!rutgers!nike!eugene
References: (numerous)

Show Full Headers
Complete Article Headers
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: arc1-ca.pugleaf.net!utzoo!mnetor!seismo!rutgers!nike!eugene
From: eugene@nike.uucp (Eugene Miya N.)
Newsgroups: net.arch
Subject: Let's start a trend! (Performance measurement -- units)
Message-ID: <722@nike.UUCP>
Date: Mon, 3-Nov-86 20:27:03 EST
Article-I.D.: nike.722
Posted: Mon Nov  3 20:27:03 1986
Date-Received: Tue, 4-Nov-86 03:21:47 EST
References: (numerous)
Sender: usenet@nike.UUCP
Reply-To: eugene@nike.UUCP (Eugene Miya N.)
Organization: NASA Ames Research Center, Moffett Field, CA.
Lines: 48
Keywords: MIPS, MFLOPS, Macho, IBM 3090