K312240@AEARN.BITNET (Klaus Kusche) (10/31/90)
Dear Mailing List: Concerning the Inmos policy of not discussing software problems on the net: Could some Inmos representative please explain the reasons for that decision??? For academics, email is most comfortable: It's free, it's fast, and every academic has a net-able terminal on his desk... Also, code pieces would be machine readable to try them immediately. Also, I don't see any legal problems: Any network allows traffic from or to commercials as long as the other end is academic (as least as far as I know). Concerning transputer task priorities and dirty/undocumented features: All legal ways of changing the priority of a transputer process (see recent postings and the CWG) involve a process switch. There is a dirty trick: Do a "gajw" (general adjust workspace) with low order bit one. The docu says that this is undefined (low order two bits must be zero), but it has been reported to silently toggle (invert) the current process' priority, from high to low or vice versa. Advantage: Fast, no process switch involved. Disadvantage: Will most likely not work on future transputers (not even tested with T801, T805 etc, does somebody know?) The hw scheduler doesn't quite get what is going on, so make sure you switch back before any scheduling occurs... (the queue's and status save areas are likely to get corrupted..) Greetings ************************************************************************ * Klaus Kusche * * Research Institute for Symbolic Computation * * Johannes Kepler University Tel: +43 7236 3231 67 * * A-4040 Linz Telex: (Austria) 22323 uni li a * * Austria (Europe) Fax: +43 7236 3338 30 * * * * Bitnet: K312240@AEARN * * Arpa/CS/Internet: K312240%AEARN.BITNET@CUNYVM.CUNY.EDU * * UUCP: mcsun!aearn.bitnet!K312240 * * Janet: k312240@earn.aearn or k312240%aearn@earn-relay * ************************************************************************