Re: SPARC code generation for compiler course

salomon@silver.cs.umanitoba.ca (Daniel J. Salomon)
Wed, 15 Dec 1993 18:18:43 GMT

          From comp.compilers

Related articles
SPARC code generation for compiler course mackey@cse.ucsc.edu (1993-12-10)
Re: SPARC code generation for compiler course salomon@silver.cs.umanitoba.ca (1993-12-13)
Re: SPARC code generation for compiler course chase@Think.COM (1993-12-13)
Re: SPARC code generation for compiler course markt@harlequin.co.uk (1993-12-14)
Re: SPARC code generation for compiler course salomon@silver.cs.umanitoba.ca (1993-12-15)
Re: SPARC code generation for compiler course pardo@cs.washington.edu (1993-12-16)
Re: SPARC code generation for compiler course nickh@harlequin.co.uk (1993-12-16)
Re: SPARC code generation for compiler course chase@Think.COM (1993-12-16)
Re: SPARC code generation for compiler course pardo@cs.washington.edu (1993-12-17)
| List of all articles for this month |
Newsgroups: comp.compilers,comp.arch
From: salomon@silver.cs.umanitoba.ca (Daniel J. Salomon)
Keywords: architecture, courses, sparc
Organization: Computer Science, University of Manitoba, Winnipeg, Canada
References: 93-12-040 93-12-053
Date: Wed, 15 Dec 1993 18:18:43 GMT

I wrote:
> One important point is not to use the program stack for temporaries.
> The SPARC expects the stack structure to be consistent at all times so
> it can save overflow register windows properly.


Many readers have misinterpreted this statement. What I was actually
warning against was using the top of the stack for expression evaluation,
like in a real stack machine. Some of my students were doing this and
always getting their stack values clobbered by register saves. The top of
the stack must be kept ready to to hold saved registers. You can still,
of course, use the stack for holding local variables, as long as they are
only semi-dynamic.
--
Daniel J. Salomon -- salomon@cs.UManitoba.CA
              Dept. of Computer Science / University of Manitoba
              Winnipeg, Manitoba, Canada R3T 2N2 / (204) 474-8687
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.