Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!tank!uxc!garcon!garcon.cso.uiuc.edu!grunwald From: grunwald@flute.cs.uiuc.edu Newsgroups: comp.arch Subject: Re: Register Scoreboarding Message-ID: Date: 9 May 89 16:34:43 GMT References: <24821@lll-winken.LLNL.GOV> Sender: news@garcon.cso.uiuc.edu Organization: University of Illinois, Urbana-Champaign Lines: 17 In-reply-to: brooks@maddog.llnl.gov's message of 7 May 89 07:47:17 GMT Hennessy recently gave a talk here & in response to a question concerning either scoreboarding or register windows, posited that object-code recompilers would be able to solve the same set of problems (binary compatibility across architectural changes), cheaper. In his picture, when you have a new architecture with, say, more registers, different delay costs or deeper pipelines, you translate your .o files and/or your final binaries. This is essentially what scoreboarding is doing, albeit dynamically. There are some limits to this approach, some obvious, some not so obvious. Comments? -- Dirk Grunwald Univ. of Illinois grunwald@flute.cs.uiuc.edu