home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.fortran
- Path: sparky!uunet!ukma!cs.widener.edu!eff!news.oc.com!convex!joelw
- From: joelw@convex.com (Joel Williamson)
- Subject: Re: loop reduction problem on Silicon Graphics
- Sender: usenet@news.eng.convex.com (news access account)
- Message-ID: <1993Jan28.174523.27370@news.eng.convex.com>
- Date: Thu, 28 Jan 1993 17:45:23 GMT
- References: <1993Jan28.112542@ltssun8.epfl.ch>
- Nntp-Posting-Host: mozart.convex.com
- Organization: Engineering, CONVEX Computer Corp., Richardson, Tx., USA
- X-Disclaimer: This message was written by a user at CONVEX Computer
- Corp. The opinions expressed are those of the user and
- not necessarily those of CONVEX.
- Lines: 23
-
- In article <1993Jan28.112542@ltssun8.epfl.ch> dubuf@ltssun7.epfl.ch (Hans du Buf) writes:
- >
- >I want to apply the same procedure in getting e.g. the maximum value in an array:
- >
- > rmax=-1.0 !assume a(i) >= 0
- >C$DOACROSS SHARE(a),LOCAL(i),REDUCTION(rmax)
- > do 10 i=1,1000
- > if (a(i).gt.rmax) rmax=a(i)
- >10 continue
- >
-
- Regardless of whether it parallelizes, you'll get better performance
- with
-
- do 10 i=1,1000
- rmax = max (a(i), rmax)
- 10 continue
-
- and it just might parallelize as well.
-
- Joel Williamson
- --
-
-