(Proposal) patch to enable RTL checks by default

Kaveh R. Ghazi ghazi@caip.rutgers.edu
Tue Oct 24 13:42:00 GMT 2000


 > From: Geoff Keating <geoffk@cygnus.com>
 > 
 > "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:
 > 
 > > I wish it was, currently we turn on "misc", "tree" and "gc" by default.
 > > Perhaps I'll propose making rtl checking on by default too.
 > 
 > I think rtl checking makes the compiler much slower (like 10x slower),
 > which is why it's not on by default.


I regularly turn on RTL checking and I typically see much less
slowdown, e.g. 50% - 75%.  Its only when memory is very low and
thrashing comes into play that I see anything like 10x (1000%.)



 > From: Mark Mitchell <mark@codesourcery.com>
 > 
 > >>>>> "Kaveh" == Kaveh R Ghazi <ghazi@caip.rutgers.edu> writes:
 > 
 >     Kaveh> Okay to install?
 > 
 > I'd rather not do that.  
 > 
 > In my experience, it's pretty expensive, even on pretty
 > high-performance machines.
 > 
 > I do think it would be reasonable to turn it on in Geoff's
 > regression-checker, or another automated system.  It's important we
 > find these problems, but I think we can do that after-the-fact, rather
 > than up front.


Many (most?) problems exposed by RTL checking are very target
specific, so only turning it on in Geoff's system isn't complete
coverage, but I guess its a start. :-)

Geoff, can your regression checker handle it?

		--Kaveh
--
Kaveh R. Ghazi			Engagement Manager / Project Services
ghazi@caip.rutgers.edu		Qwest Internet Solutions


More information about the Gcc-patches mailing list