Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!oliveb!sun!shukra!ram
From: ram%shukra@Sun.COM (Renu Raman, Taco-Bell Microsystems)
Newsgroups: comp.arch
Subject: Re: 2-D arrays
Message-ID: <44258@sun.uucp>
Date: 4 Mar 88 18:41:39 GMT
References: <179@wsccs.UUCP: <696@nuchat.UUCP> <284@scdpyr.UUCP> <25699@linus.UUCP> <2378@pdn.UUCP> <1020@PT.CS.CMU.EDU> <244@eos.UUCP>
Sender: news@sun.uucp
Reply-To: ram@sun.UUCP (Renu Raman, Taco-Bell Microsystems)
Organization: Sun Microsystems, Mountain View
Lines: 31

In article <244@eos.UUCP> eugene@eos.UUCP (Eugene Miya) writes:
>A defect?  I'll pass that point of view on to the chief architect of
>the machine the next time I see him (two weeks).  A feature! 8-)

  Is it D.J. Kuck or Stokes?

  In any case, in the BSP, Kuck et. al. designed a memory system where the
  interleave size and bank size were relatively prime (16 & 17 I think. - ref:
  paper by kuck & Stokes on BSP in IEEE TOC, May 82.  You can probably find
  it in Faye's book too).

  Only for succesive access to an array  that had a displacement of 17,
  would there be conflicts and I doubt if 17 was a favourite of
  Fortran Programmers.  

  The design is complicated by the Alignment network that is needed (you
  can find all the gory details in one of those refs.).  An interesting
  application of this is in Frame Buffer design (more so with
  the multi-processor graphics engines) and probably for LIW machines.

>A good paper, BTW, is Duncan Lawrie's "Access and Alignment of Data" paper
>in IEEE TOC 1971.  But he's over at the Cathedral Hill at COMPCON right now.
>
>From the Rock of Ages Home for Retired Hackers:
>
>--eugene miya, NASA Ames Research Center, eugene@ames-aurora.ARPA

---------------------
   Renukanthan Raman				ARPA:ram@sun.com
   Sun Microsystems			UUCP:{ucbvax,seismo,hplabs}!sun!ram
   M/S 5-40, 2500 Garcia Avenue,
   Mt. View,  CA 94043