Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!cam-eng!tpm From: tpm@eng.cam.ac.uk (tim marsland) Newsgroups: comp.sys.transputer Subject: Re: C++ on Transputer Keywords: Translator, Compiler, C++, who Message-ID: <1130@rasp.eng.cam.ac.uk> Date: 10 Feb 90 18:02:02 GMT References: <362@b21.UUCP> Sender: tpm@eng.cam.ac.uk Reply-To: tpm@eng.cam.ac.uk (tim marsland) Organization: Cambridge University Engineering Department, UK Lines: 32 In article <362@b21.UUCP> lohnert@b21.UUCP (Frieder Lohnert) writes: > I am looking for C++ Translator/Compiler for Transputers. > Who is selling or planning such a product ? i believe Cogent Research were using c++ ``in anger'' on transputers two or more years ago. according to a recent glossy i saw, their current XTM machine runs c, c++ and fortran. > Is there in any experience in porting C++ Translator/Compiler for the > Transputer ? re: porting, almost certainly, this has been done before in many places. our experiences have been that "porting" the at&t cfront is really quite straightforward provided your c compiler behaves. About 18 months ago a research student in our group (who had written a lot of c++ code) used c++ to implement a distributed hmm on a ring of 64 transputers. He was able (well more or less) to run the output of the 1.1 cfront translator running on under ultrix, into the logical systems transputer toolset c cross-compiler running under sunos (did anyone say contorted?). At that time, we simply sneaked around the static constructor nm/munch bits :-) i've currently got a project student working on applications of dynamic switching, and as a prelude to that, he's doing the port of 1.2 cfront plus libraries to the 88.4 version of the logical systems c compiler -- all seems fine so far. I've thrown together a restricted version of nm(1) which trundles through logical systems object files and libraries and talks to the standard munch in the right kind of way, so we're going to try and do static constructors 'properly' this time. tim marsland