home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.parallel
- Path: sparky!uunet!gatech!hubcap!fpst
- From: kunisada@is.kyushu-u.ac.jp (Katuhiro KUNISADA)
- Subject: Re: definition of parallel efficiency
- In-Reply-To: kitchel@manta.dpsi.com's message of Mon, 11 Jan 1993 15:28:55 GMT
- X-Nntp-Posting-Host: is-og8.is.kyushu-u.ac.jp
- Message-ID: <KUNISADA.93Jan22162745@is-og8.is.kyushu-u.ac.jp>
- Sender: news@is.kyushu-u.ac.jp
- Organization: Dept. of Information Systems, Kyushu Univ., Japan
- References: <1993Jan11.142638.7789@hubcap.clemson.edu>
- <kitchel.726766135@manta>
- Date: Fri, 22 Jan 1993 07:27:45 GMT
- Approved: parallel@hubcap.clemson.edu
- Lines: 44
-
- In article <kitchel.726766135@manta> kitchel@manta.dpsi.com (Sidney W. Kitchel) writes:
-
- >Stephen Vavasis <vavasis@cs.cornell.EDU> writes:
- >
- >
- >>I have seen two definitions of "parallel efficiency" in the literature,
- >>and I am wondering whether one definition or the other has prevailed.
- >>They are:
- >
- >>(1) par eff = (seq. time of the par. algorithm) / (p * parallel time)
- >
- >>(2) par eff = (seq. time of the best possible sequential algorithm
- >> for the problem) / (p * parallel time)
- >
- >>If you had a parallel cyclic reduction algorithm, would you rate it
- >>against a sequential cyclic reduction algorithm or against tridiagonal
- >>Cholesky factorization?
- >
- >>How about when the underlying problem is more complicated, for example
- >>solving a boundary value problem? In this case it's not clear how to
- >>identify the best possible sequential algorithm.
- >
- >>I would like to get this definition straight because I am teaching a
- >>class on parallelism.
- >
- > Since Frontiers '92, it also depends on whether you did the
- >programming or God did.
-
-
- As we are not god.
-
- I think that, the best way to estimate parallel efficiency is to use
- the sequential algorithm that is known to perform best, as the following.
-
- (worst seq. time of the seq. algorithm that is known to be best)
- par eff = ----------------------------------------------------------------
- (p * parallel time of the worst case)
-
-
-
- ----------------------------------------------------------------------
- K. kunisada kunisada@is.kyushu-u.ac.jp
- My name is same with the most famous Yakuza (Kunisada Tyuuji)
-
-