Vectorising compilers
Peter W. Brewer
lerici at super.ORG
Wed Aug 31 08:43:52 AEST 1988
In article <13265 at mimsy.UUCP> chris at mimsy.UUCP (Chris Torek) writes:
>In article <1203 at radio.toronto.edu> brian at radio.astro.toronto.edu
>(Brian Glendenning) writes:
>>Are vectorizing C++ compilers available on "crunching" machines, e.g.
>>Cray, Convex and Alliant? (In fact, are vectorizing _C_ compilers
>>available for the latter two)?
>
>Convex has (and has had for years) a vectorising C compiler. It is the
>same as their FORTRAN back-end. And if you use a Pascal-to-C translator,
>you even get vectorised Pascal!...
>--
>In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163)
>Domain: chris at mimsy.umd.edu Path: uunet!mimsy!chris
I have ported the AT&T version to the Convex C-1 and with alterations to allow
the Convex long long 64-bit type. It works with the Convex vectorizing C compiler
and with the Convex generic C compiler. Both of these have been extended to allow
the 64-bit integer type long long. The Convex vectorizing C-compiler is also a
very good optimizing scalar compiler.. however there are, as usual, some problems
when it compiles things like YACC generated stuff even when all optimizations are
turned off. And so for various reasons my rendition of cfront on the Convex has
to use both versions of Convex C compiler technology. And so yes there is a vectorizing
C++ on the convex... soon I hope to port it to a Cray 2... as soon as Cray gives
us a working C compiler {It is still too buggy and it is going to be difficult to
port to a machine which is not byte addressable.. is not really System V even though
it claims to be and yet is not Berkeley.. :) shades of SUN OS 4.0! (: The version of
the Convex C++ is still undergoing testing. I can compile a working cfront with it.
But it seems to me someone should develop a test suite in order to find out if things
like C++ or even vectorizing/optimizing C's do the right thing. Any takers?
Peter Brewer
lerici at super.org
--
Peter Brewer |_____________| THE
lerici at super.org |___|____|/ SUPERCOMPUTING
|__ |__ |_/ RESEARCH
|___|__ /_| CENTER
More information about the Comp.lang.c
mailing list