Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!decwrl!sun-barr!newstop!texsun!convex!metzger From: metzger@convex.com (Robert Metzger) Newsgroups: comp.benchmarks Subject: Re: benchmarks and interprocedural optimization Message-ID: <1991Mar21.212212.18139@convex.com> Date: 21 Mar 91 21:22:12 GMT References: <1991Mar21.184428.8226@convex.com> Sender: news@convex.com (news access account) Organization: Convex Computer Corporation, Richardson, Tx. Lines: 21 Nntp-Posting-Host: bach.convex.com In article mccalpin@perelandra.cms.udel.edu (John D. McCalpin) writes: > >I was told by a source within Convex that the Application Compiler did >not, in fact, manage to eliminate all of the unnecessary checks in the >level-1 BLAS routines used in the LINPACK code. The reason had >something to do with being able to optimize expressions after the >inlining, but not logical predicates -- though I never really >understood the explanation. > >Is there a "definitive answer" from Convex on this? There is a least one useless test in LINPACK which requires predicate (eg. n>=1&&n<=100) propagation to delete. The Application Compiler performs constant propagation (eg. n==100) in V1.0, but not generalized predicate propagation. Sooner or later it will. -- Robert Metzger CONVEX Computer Corp. Richardson, Texas Generic Disclaimer: I write software, not CONVEX Corporate Policies. "The only legitimate function of government is to protect its citizens from harm or property loss by violence, stealth, or fraud. All else is tyranny."