stt@inmet.inmet.com (11/29/90)
Re: time spent in GCC's "parse" phase.
I would assume that this "parse" phase is actually the entire front-end.
I doubt if they separate out the time actually spent walking the parse
tables themselves. I agree with others that this aspect is not a big
deal. Turbo C and Think C probably get their speed by doing most of the
front-end processing while the user is typing in the program.
S. Tucker Taft
Intermetrics, Inc.
Cambridge, MA 02138
[Turbo is a conventional compiler that compiles regular old ASCII source files
into binary code. It buffers entire files in memory, compiles directly to
object code without a separate assembler step, and uses a very fast and
simple linker. Nothing exotic, just good engineering. -John]
--
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.moss@cs.umass.edu (12/02/90)
>>>>> On 28 Nov 90 19:55:00 GMT, stt@inmet.inmet.com said:
Taft> Re: time spent in GCC's "parse" phase.
Taft> ... Turbo C and Think C probably get their speed by doing most of the
Taft> front-end processing while the user is typing in the program.
I don't think so. The compiler runs separately from the editor and really does
read all the source file(s) from disk. I think they get most of their speed
by hand coding a lot of crucial routines in assembly, and by focusing on speed
in designing their compilers. My guess is that they feel it is a market very
sensitive to this issue, and hence worth the effort. (It's also worth the
effort because of the profit and volume.)
--
J. Eliot B. Moss, Assistant Professor
Department of Computer and Information Science
Lederle Graduate Research Center
University of Massachusetts
Amherst, MA 01003
(413) 545-4206; Moss@cs.umass.edu
--
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.