Path: utzoo!utgpu!water!watmath!ccplumb
From: ccplumb@watmath.waterloo.edu (Colin Plumb)
Newsgroups: comp.arch
Subject: Re: Memory bank conflicts
Message-ID: <17459@watmath.waterloo.edu>
Date: 14 Mar 88 00:18:55 GMT
References: <7690@pur-ee.UUCP> <3300021@uiucdcsm> <4712@lll-winken.llnl.gov> <12514@sgi.SGI.COM> <5863@ames.arpa>
Reply-To: ccplumb@watmath.waterloo.edu (Colin Plumb)
Organization: U of Waterloo, Ontario
Lines: 19
Confusion: U. of Waterloo, Ontario

eugene@pioneer.UUCP (Eugene N. Miya) wrote:
>I disagree with your opinion about separating algorithms from architectures
>from the standpoint of studying them, but I have to agree that posting
>pure algorithm stuff in the architecture group is wasting bits.

Well, as others have observed, optimizing compilers are an integral
part of RISC architectures, VLIW architectures, vector architectures,...

I think you can go quite deep into software algorithms without getting
beyond comp.arch's purview.  Certainly the discussion of reorganization
algorithms is applicable.

So bring on the killer optimizing compiler technology!  There's no such
thing as too fast.
--
	-Colin (watmath!ccplumb)

Zippy says:
- if it GLISTENS, gobble it!!