Path: utzoo!attcan!uunet!lll-winken!lll-ncis!helios.ee.lbl.gov!pasteur!ames!hc!pprg.unm.edu!unmvax!gatech!gitpyr!loligo!mccalpin
From: mccalpin@loligo.cc.fsu.edu (John McCalpin)
Newsgroups: comp.arch
Subject: IEEE floating-point on supercomputers
Keywords: multiple-precision floating-point, IEEE standard floating-point
Message-ID: <366@loligo.cc.fsu.edu>
Date: 5 Jan 89 20:01:06 GMT
References: <325@loligo.fsu.edu> <20295@ames.arc.nasa.gov>
Reply-To: mccalpin@masig1.ocean.fsu.edu (John McCalpin)
Organization: Supercomputer Computations Research Institute
Lines: 36

In article <20295@ames.arc.nasa.gov> (Hugh LaMaster) writes:
>In article <325@loligo.fsu.edu> (John D. McCalpin) writes:

>>There is some hesitancy in the supercomputer community to switch to the
>>IEEE format because the exponent range of 64-bit numbers is so much
>>smaller than the range currently provided by Cray and CDC/ETA formats.

>Most users that I know
>of would welcome having the same format on their supercomputer as on
>their graphics engine (often IEEE), and the IEEE format is considered
>among the best available by numerical analysts.  ("welcome" seems a little
>weak in retrospect.  Some people would kill for it.  

>  Hugh LaMaster, m/s 233-9,  UUCP ames!lamaster
>  NASA Ames Research Center  ARPA lamaster@ames.arc.nasa.gov

I certainly count myself in the group that longs dearly for binary
data portability.  I complain to my contacts at CDC at every available
opportunity, and ask: "WHY has no supercomputer vender EVER supplied
a front-end which uses the same floating-point format and instruction
set as the back end?"  The response is always some variation on: "Who
is going to pay for it?"

I suspect that one of the reasons that the Cyber 205 never caught on is
that (because of the inability to move data) everyone expects their
supercomputer to do EVERYTHING well -- including all the scalar stuff
that could be done far more cost-effectively on a front-end.  Cray has
not (to my mind) solved the problem, they have just got a faster scalar
machine....  I am still waiting for a real binary-compatible
distributed system which includes a supercomputer - maybe DEC needs to
come out with a vector-pipelined VAX supercomputer.... :-)
-- 
----------------------   John D. McCalpin   ------------------------
Dept of Oceanography & Supercomputer Computations Research Institute
mccalpin@masig1.ocean.fsu.edu		mccalpin@nu.cs.fsu.edu
------------------------------------------------------------------