From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,e29c511c2b08561c X-Google-Attributes: gid103376,public From: dewar@cs.nyu.edu (Robert Dewar) Subject: Re: Is the "Ada mandate" being reconsidered? Date: 1996/06/13 Message-ID: #1/1 X-Deja-AN: 160480908 references: <4mq7mg$8hs@jake.probe.net> <4peu0v$rfq@news15.erols.com> <1996Jun10.114827.26046@relay.nswc.navy.mil> <4pk5sm$i7k@gde.GDEsystems.COM> <4pn0rs$mbe@gde.GDEsystems.COM> organization: Courant Institute of Mathematical Sciences newsgroups: comp.lang.ada Date: 1996-06-13T00:00:00+00:00 List-Id: Tom Robinson says "I am interested in what measure you are using for quality. From the small snipits I have read on the net it is not clear that the generated code quality of gnat is quite up to current standards yet. Am I missing something here? Has anyone done ACEC or PIWG comparisons yet?" As for most compilers, GNAT does some things well, and some things not as well. For some programs, GNAT is the fastest Ada compiler around (in terms of generated code), these are programs for which the backend of GCC (which is very highly optimized) can generate really good code. For some Ada specific constructs (notably check generation, aggregates, exception handling and unconstrained array handling), there is a way to go, but look to see major improvements in GNAT in all these areas in the near future. If you benchmark GNAT against other compilers, you will find that some things are faster, some are slower. Overall, the results are pretty comparable, which is quite interesting considering that in these cases you are typically comparing an old technology where a LOT of effort has been put into improving the quality of code in general (and the PIWG's in particular) with a new technology where so far almost no effort has been put into improving efficiency of generated code (we have been much more concerned with functionality and reliability). Robert Dewar ACT