Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!think.com!news From: rlk@think.com (Robert Krawitz) Newsgroups: comp.arch Subject: Re: massive parallelism, was CDC 6600 and TI ASC Message-ID: <1991Mar11.220703.13544@Think.COM> Date: 11 Mar 91 22:07:03 GMT References: <45252@ut-emx.uucp> <1991Mar7.215545.430@zoo.toronto.edu> <780@llnl.LLNL.GOV> Sender: news@Think.COM Reply-To: rlk@think.com (Robert Krawitz) Organization: Thinking Machines Corp., Cambridge MA Lines: 18 In-Reply-To: brooks@physics.llnl.gov (Eugene D. Brooks III) In article <780@llnl.LLNL.GOV>, brooks@physics (Eugene D. Brooks III) writes: ]In article lamson@el1.crd.ge.com (scott h lamson) writes: ]Yes, it will fail for the same historical reasons. ]NASA AMES has some impressive computational efficiencies for this ]machine on 3-D FFTs, the number I saw on a viewgraph was "~0" This timing was performed in 64-bit precision on a machine with 32 bit floating point hardware (effectively, without floating point hardware). We do offer 64 bit floating point hardware. The timing was also for considerably more than just a 3D FFT. The performance for the FFT's in question, using currently-released software and 64 bit FPU hardware, is between 45 and 140 Mflops, depending upon the options chosen, on an 8K processor machine (1/8 of a full machine). -- ames >>>>>>>>> | Robert Krawitz 245 First St. bloom-beacon > |think!rlk (postmaster) Cambridge, MA 02142 harvard >>>>>> . Thinking Machines Corp. (617)234-2116