🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.intel
Article #1262

Re: Superscalar vs. multiple CPUs ?

#1262
From: Stephen Spackman
Date: Sun, 20 Dec 1992 14:31
30 lines
2606 bytes
No; actually the real issue is whether PCs and workstations in fact run
batch mode. They don't. Contrary to what the market was mysteriously
made to swallow with the IBM PC and the Macintosh, what I do on my
workstation is in no way single threaded - at very *least* I have I/O
running in parallel with my application kernel (and in fact I usually
have a newsreader, a mailer, an editor (that's three copies of epoch
[emacs], actually), some language implementation, and half a dozen
monitor programmes (xloads, clocks, biffs, faces...) running. And that's
on a mere sparc....

If we start doing intelligent things like parallel garbage collection,
self-reorganising filesystems, code-generation-as-caching, we can
probably get a dozen processes running that have no desire to sleep.

The general-purpose world just isn't single-threaded. Nor, for that
matter, is it fp-intesive. But marketting determines reality, of course,
and people *will* keep insisting on engineering PCs for flops rather
than editable filesize, network transparency and user interface
responsiveness!

But no, the real reason is that the hardware and the software are never
optimised for each other; they merely co-evolve. So we are left with
hardware Unix engines and hardware DOS engines and C the best language
to programme them in. Probably, there is nothing we can do to change
that, so probably any idea related to micro-task-switching is doomed to
failure in the marketplace.
----------------------------------------------------------------------
stephen p spackman    +49 681 302 5288(o) 5282(sec)    stephen@acm.org
      dfki / stuhlsatzenhausweg 3 / d-w-6600 saarbruecken 11 / germany
----------------------------------------------------------------------
Technical Details
Message-ID: <SPACKMAN.92Dec20153527@disco-sol.dfki.uni-sb.de>
Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!spool.mu.edu!yale.edu!ira.uka.de!sbusol.rz.uni-sb.de!coli.uni-sb.de!coli-gate.coli.uni-sb.de!spackman
References: <1992Dec7.012026.11482@athena.mit.edu> <1992Dec8.000357.26577@newsroom.utas.edu.au> <PCG.92Dec9154602@aberdb.aber.ac.uk> <1992Dec12.152224.168173@zeus.calpoly.edu> <PCG.92Dec19165907@decb.aber.ac.uk>

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!spool.mu.edu!yale.edu!ira.uka.de!sbusol.rz.uni-sb.de!coli.uni-sb.de!coli-gate.coli.uni-sb.de!spackman
From: spackman@disco-sol.dfki.uni-sb.de (Stephen Spackman)
Newsgroups: comp.sys.intel,comp.arch
Subject: Re: Superscalar vs. multiple CPUs ?
Date: 20 Dec 1992 14:31:12 GMT
Organization: DFKI Saarbruecken GmbH, D-W 6600 Saarbruecken
Lines: 30
Message-ID: <SPACKMAN.92Dec20153527@disco-sol.dfki.uni-sb.de>
References: <1992Dec7.012026.11482@athena.mit.edu>
 <1992Dec8.000357.26577@newsroom.utas.edu.au>
 <PCG.92Dec9154602@aberdb.aber.ac.uk>
 <1992Dec12.152224.168173@zeus.calpoly.edu>
 <PCG.92Dec19165907@decb.aber.ac.uk>
Reply-To: stephen@acm.org
NNTP-Posting-Host: disco-sol.dfki.uni-sb.de
In-reply-to: pcg@aber.ac.uk's message of 19 Dec 92 16:59:07 GMT