Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!bloom-beacon!mit-eddie!uw-beaver!cornell!batcomputer!itsgw!steinmetz!uunet!mcvax!ruuinf!atze From: atze@ruuinf.UUCP (Drs. A. Dijkstra) Newsgroups: comp.sys.transputer Subject: Re: PARSEC's parallel C compiler Par.C ?!?! Summary: experience with PARSEC's C-compiler Message-ID: <502@ruuinf.UUCP> Date: 21 Jun 88 14:34:35 GMT References: <8806170829.AA22577@uk.ac.ox.prg> Organization: Univ of Utrecht, Dept of CS Lines: 60 In article <8806170829.AA22577@uk.ac.ox.prg>, PVR%bgerug51.earn@NSS.CS.UCL.AC.UK writes: > Dear mailing list, > > I had an interesting discussion wednesday at the Transputer symposium in > Utrecht with someone of the Dutch company PARSEC (not Parsytec !), who's > parallel C compiler is now supported by Parsytec (not Parsec !). In their > parallel C compiler Par.C the parallel features of occam have been embedded in > the language (par, guarded alt, channel type, ...). Par.C is compatible with > K&R, with the extensions of H&S. It also translates a number of keywords. > stuff deleted > 1) is there anyone using this promising C compiler yet ? I have been using a testversion (oct-87) of the C-compiler in a project involving weather forecasting on transputers. Originally the compiler runs on a PC, but an easy port was made to let the compiler run on a UNIX system (HCX/9, Harris). The compiler at that time contained a lot of bugs w.r.t. the special OCCAM-alike features (bugs are reported to PARSEC), but it was possible to program around them and to produce fairly large parallel programs, which also worked. Support software existed of stdio routines, but only for the transputer connected to the host and some math routines. Software for booting the network had to be written by myself but was largely done using the C-compiler. Sofar the bad news. The good news is that 1 july 1988 a new version will appear, which will be changed (not upgraded) to a version which is probably more final than the version I worked with (call given address). I have no experience with the new version of the compiler yet, but discussions with developers from PARSEC are promising. Apart from all this information, the builtin parallel/alternative/timer constructs were nice to use, implemented quite fast and compared with the OCCAM TDS system it was a relief to use. C is much more flexible and developing on a Unix system is also much more convenient > > -------------------------------------- > Parsec Developments can be contacted @ > Witte Singel 66, > 2311 BM Leiden, > Phone: +31 71 142 142 > TeleFax:+31 71 134 449 > > Patrick ** Atze Dijkstra ** ** Dept. of CS, RUU ** ** Postbus 80.089 ** ** 3508 TB Utrecht ** ** Padualaan 14 ** ** 3584 CH Utrecht ** ** The Netherlands ** ** tel. +31-30-531935/531454 ** ** EMail: UUCP : .. mcvax!ruuinf!atze **