keith@uw-apl.UUCP (Keith Kerr) (07/27/89)
Is there an implementation of C++ running on cray./UNICOS systems?? If not, could anyone venture a guess as to how tough a port would be. I've heard that UNICOS was derived from some unix version, but I don't know anything more about it.
mjt@super.ORG (Michael J. Tighe) (08/01/89)
In article <241@uw-apl.UUCP> keith@uw-apl.UUCP (Keith Kerr) writes: > Is there an implementation of C++ running on cray./UNICOS systems?? If > not, could anyone venture a guess as to how tough a port would be. > I've heard that UNICOS was derived from some unix version, but I don't > know anything more about it. I do not know of any C++ implementations for UNICOS. I doubt CRI will be coming out with one anytime soon. The difficulty factor depends on how you plan to go about it. If you plan to write it from the ground up, I would say it would be very difficult (unless of course, you have done this sort of thing before). If you plan to write it as a preprocessor, that would be somewhat easier. UNICOS was originally a port of a version of Unix that ran on a Vax. ------------- Michael Tighe internet: mjt@super.org uunet: ...!uunet!super!mjt -- ------------- Michael Tighe internet: mjt@super.org uunet: ...!uunet!super!mjt
rosenkra@hall.cray.com (Bill Rosenkranz) (08/01/89)
In article <241@uw-apl.UUCP> keith@uw-apl.UUCP (Keith Kerr) writes:
=Is there an implementation of C++ running on cray./UNICOS
=systems?? If not, could anyone venture a guess as to
=how tough a port would be. I've heard that UNICOS was
=derived from some unix version, but I don't know anything
=more about it.
UNICOS is based on UNIX System V.3 so in theory, cfront should be possible.
i am not an expert on c++, but as i understand it, the research version
(1.2.1?) is a preprocessor which generates c code. cray has an ANSI-
compliant C as well as a vectorizing C which so far i have been able
to port numerous codes with (i have just started a port of GNU emacs
18.54 for a cray-2 under UNICOS 5.0, for example). so c++ is text in,
text out and it should not be a problem. the question is whether the
c code generated can readily be compiled. i suspect it can.
i have no knowledge of cray's plans for c++ but i suspect somebody
is playing around with it :^).
-bill
rosenkra@boston.cray.com
[i speak for me, not cray, so don't quote me...]
malcolm@Apple.COM (Malcolm Slaney) (08/01/89)
In article <12296@super.ORG> mjt@super.UUCP (Michael J. Tighe) writes: >I do not know of any C++ implementations for UNICOS. I doubt CRI will >be coming out with one anytime soon. The difficulty factor depends on >how you plan to go about it. I don't agree. I've attended the last couple of Cray C Advisory Board meetings and the C group wants to know how much interest there is in C++. If enough people want it then it will become a product. C++ is just a preprocessor. In fact it probably runs best on a Unix System V machine just like Unicos. Call up your local AT&T salescritter and ask them for a source tape. (Didn't AT&T just announce a the new release with multiple inheritance? I wouldn't be surprised if this already compiles and runs on the Cray.) So here's the offer. If you want C++ on the Cray send me a note and I'll forward them all to the head of the C compiler group at Cray. (I don't like publishing other people's addresses.) Cray, its just another generic Unix box.....but boy do I love those Flops. Malcolm
mjt@super.ORG (Michael J. Tighe) (08/02/89)
In article <33634@apple.Apple.COM> malcolm@Apple.COM (Malcolm Slaney) writes: >> In article <12296@super.ORG> mjt@super.UUCP (Michael J. Tighe) writes: >> I do not know of any C++ implementations for UNICOS. I doubt CRI will >> be coming out with one anytime soon. > I don't agree. I've attended the last couple of Cray C Advisory Board > meetings and the C group wants to know how much interest there is in > C++. If enough people want it then it will become a product. Well Malcolm, you can disagree, but I don't know of any C++ implementations for UNICOS. Really. Trust me. I take it as a complement that you forgot I was sitting next to you at the last Cray C Advisory Board meeting in May. I distinctly recall (and have it in my notes) that Cray asked for feedback on C++. We all agreed that "since there are many problems associated with C++, the best move would be to wait." Given this, I think it is fair to say it won't be coming out anytime soon. Mike -- ------------- Michael Tighe internet: mjt@super.org uunet: ...!uunet!super!mjt
jac@muslix.llnl.gov (James Crotinger) (08/10/89)
The Livermore Computer Center (LCC) at Lawrence Livermore National Lab has been working on a port of CFRONT 1.2 to run under either LTSS or NLTSS (not sure which). It is to be used as a front end for their Hybrid C Compiler (HCC), which is an ANSI conforming C compiler (or about as close as you can come). I know that LTSS is moving towards "UNIX compliance" by providing UNIX compatible libraries and system calls, so moving their port to UNICOS might not be too difficult. Jim
fouts@garth.ingr.com (Martin Fouts) (08/17/89)
In article <3930@hall.cray.com> rosenkra@hall.cray.com (Bill Rosenkranz) writes: In article <241@uw-apl.UUCP> keith@uw-apl.UUCP (Keith Kerr) writes: =Is there an implementation of C++ running on cray./UNICOS =systems?? If not, could anyone venture a guess as to =how tough a port would be. I've heard that UNICOS was =derived from some unix version, but I don't know anything =more about it. UNICOS is based on UNIX System V.3 so in theory, cfront should be possible. i am not an expert on c++, but as i understand it, the research version (1.2.1?) is a preprocessor which generates c code. cray has an ANSI- compliant C as well as a vectorizing C which so far i have been able to port numerous codes with (i have just started a port of GNU emacs 18.54 for a cray-2 under UNICOS 5.0, for example). so c++ is text in, text out and it should not be a problem. the question is whether the c code generated can readily be compiled. i suspect it can. i have no knowledge of cray's plans for c++ but i suspect somebody is playing around with it :^). -bill rosenkra@boston.cray.com [i speak for me, not cray, so don't quote me...] Cfront should be easy to port, barring an potential number of word size and alignment problems. The problem isn't the preprocessor, it is the library support. This would be much more interesting. If you've got experience providing run time libraries and you understand UniCos, the calling sequence, and CAL, it should be a straightforward but time consuming effort. The other alternative is G++, which requires Gnu C. I started a Gnu C port to the Cray 2, and it looks like a lot of work because of the way in which Crays differ from the model provided by the Gnu compiler, so I would recommend the first approach. Marty -- Martin Fouts UUCP: ...!pyramid!garth!fouts ARPA: garth!fouts@ingr.com PHONE: (415) 852-2310 FAX: (415) 856-9224 MAIL: 2400 Geng Road, Palo Alto, CA, 94303 I'll say something clever, as soon as I get some caffeine in my blood stream -- Martin Fouts UUCP: ...!pyramid!garth!fouts ARPA: garth!fouts@ingr.com PHONE: (415) 852-2310 FAX: (415) 856-9224 MAIL: 2400 Geng Road, Palo Alto, CA, 94303 I'll say something clever, as soon as I get some caffeine in my blood stream