Path: utzoo!utgpu!watmath!clyde!ima!compilers-sender From: grow@druhi.att.com Newsgroups: comp.compilers Subject: Re: code generator-generators Summary: Two examples Message-ID: <2971@ima.ima.isc.com> Date: 1 Dec 88 20:36:16 GMT References: <2964@ima.ima.isc.com> Sender: compilers-sender@ima.ima.isc.com Reply-To: grow@druhi.att.com Lines: 31 Approved: compilers@ima.UUCP In article <2964@ima.ima.isc.com>, markhall@pyrps5.pyramid.com (Mark Hall) writes: > I'm interested in knowing whether any of the common code > generator-generators (Glanville, Cattell, Ganapathi, etc) are used in > any production compilers. ... The AIE Ada compiler done by Intermetrics Incorporated of Cambridge Mass. used the PQCC (Production Quality Compiler Compiler) technology. This compiler does very sexy instruction selections and register allocations. If anybody mentions this compiler they are very libale to say that it was a real dog as far as its speed of compilation was concerned. This is probabily true but, the blame should be placed on the intermediate representations (Diana to be exact) rather than the code generator technology! The ALS Ada compiler done the Softech Corporation of Waltham Mass. used the Graham-Glanville technology. This compiler's code wasn't nearly as sexy as the code produced by the AIE but, it was supposed to be fairly good. The same thing can be said about the speed of compilation that was said about the AIE. Note, I believe that the ALS compiler is still under development, so any comments made about it must be taken as relevant only to it's early history! -- Gary Oblock Compiler Consultant to Bell Labs att!druhi!grow (303)538-4169 -- Send compilers articles to ima!compilers or, in a pinch, to Levine@YALE.EDU Plausible paths are { decvax | harvard | yale | bbn}!ima Please send responses to the originator of the message -- I cannot forward mail accidentally sent back to compilers. Meta-mail to ima!compilers-request