home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!jvnc.net!rutgers!ub!galileo.cc.rochester.edu!joker.optics.rochester.edu!faklis
- From: faklis@joker.optics.rochester.edu (Dean Faklis)
- Newsgroups: comp.sys.ibm.pc.hardware
- Subject: Actual FLOP performance of math coprocessors
- Message-ID: <1992Aug29.193529.2694@galileo.cc.rochester.edu>
- Date: 29 Aug 92 19:35:29 GMT
- References: <94928@bu.edu>
- Sender: news@galileo.cc.rochester.edu
- Reply-To: faklis@joker.optics.rochester.edu (Dean Faklis)
- Organization: The Institute of Optics, University of Rochester
- Lines: 14
- Nntp-Posting-Host: joker.optics.rochester.edu
-
-
- I ran two tests to try and empirically determine the computational
- speed of math coprocessors.
- Using assembly language or Turbo Pascal I obtain about 500,000
- floating point operations per second. Microsoft C doesn't do any
- better.
- Does anybody have any comments on why this seems so low? All I
- was doing was some adds and multiplies?
- ( The system was a 486 @ 25 MHz)
-
- Thanks in advance for any comments
-
- Dean Faklis
- faklis@joker.optics.rochester.edu
-