Re: gcc vs. Greenhills vs. MRI

wgstuken@immd4.informatik.uni-erlangen.de (Wolfgang Stukenbrock)
Tue, 29 Oct 1991 14:13:15 GMT

          From comp.compilers
Related articles
gcc vs. Greenhills vs. MRI miklg@steinway.acuson.com (1991-10-28)
Re: gcc vs. Greenhills vs. MRI wgstuken@immd4.informatik.uni-erlangen.de (1991-10-29)
| List of all articles for this month |
Newsgroups: comp.compilers
From: wgstuken@immd4.informatik.uni-erlangen.de (Wolfgang Stukenbrock)
Keywords: C, performance, question, 68000, GCC
Organization: CSD., University of Erlangen, Germany
References: 91-10-110
Date: Tue, 29 Oct 1991 14:13:15 GMT

Be careful. I know a MRI C-compiler, that isn't able to parse correct C
programms. I know a Greenhills C-compiler, that cannot handle big
structures and arrays.


The only one (of the threee compilers you ask about) I don't know any
problems like this, is gcc.


So, test the compiler you choose, not only with a little "Hello world"
programm. Take real stuff!


At our place gcc is running cross for sun3 sun4 and sequent-symmetry.
(I've just not the time to install it for HP 300/700/800 - if supported by
gcc (700/800 I'm not sure ...)). The problem in installing gcc for a cross
environment is, that there is not enought support in it in order to do it
with a compilation. You will get wrong include paths, libary-search-path
and default-binry-search-paths. That there are incompatibilities with some
headerfiles is normal, I think.


A litle drawback of gcc. gcc is fine, but the gas is up to now not as fine
as gcc. It failed for sun4 and cross for our sequent. (I took the source
from the machines and ported for cross-installation.) Same with the gnu-ld
and the shared libaries on SUN's.


Wolfgang
--

Post a followup to this message

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