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,d207bdc3561b595 X-Google-Attributes: gid103376,public From: dewar@merv.cs.nyu.edu (Robert Dewar) Subject: Re: Realtime Extensions for Windows NT Date: 1997/05/25 Message-ID: #1/1 X-Deja-AN: 243784540 References: <3385CAC3.3EEE@eideticscorp.com> Organization: New York University Newsgroups: comp.lang.ada Date: 1997-05-25T00:00:00+00:00 List-Id: Matthew says <> I assume you are interested in Ada on these systems (or why would you be posting here?) Given this, one thing to realize about determinacy is that the Ada tasking model is entirely deterministic if Annex D is fully and completely supported. Check validation VSR's carefully to ensure that all tests that are relevant in annex D are passed. There are a few annex D tests which are legitimately optional (notably the ones for Asynchronous Task Control), but these are the only ones that should be missing. If the implementation passes 100% of these tests, you have a pretty good indication that the underlying task control system is acceptably deterministic. This is a fairly strenuous requirement. In our experience, many operating systems level thread support packages do NOT meet these requirements. For example, with the Solaris implementation of GNAT, we find that the Solaris system threads do not quite meet the requirements, even if you run in realtime mode in super user state. To pass all the annex D tests, we had to use our own threads package. A future version of the Solaris GNAT will come packaged with a choice between the two threads support options. Another possibility for you to consider is Linux operating with the FSU threads package, which passes 100% of the annex D tests.