Re: Is -fnew-ra ready for real testing?

[Michael Matz <matzmich at cs dot tu-berlin dot de>]

Hi,

On Fri, 19 Jul 2002, Brad Lucier wrote:

> I'm trying to build a beta Gambit-C with mainline gcc and -fnew-ra.
> So far, I'm getting > 1 hour compile times and multigigabyte space
> requirements for sparcv9 -m64 for some binaries.

Ahh, these problems again ;-)  And without -fnew-ra you don't see this?
In that past I sometimes tested the allocator on some of your ugly
testcases, but that was some months ago, so I possibly have screwed
performance again.

> Should I just wait to try to test-stress this thing?

I'm not sure, if the allocator in mainline will change too much, partly
because the feature freeze.  Bug fixes and smaller adjustments will be
made, though, so you can just send me a preprocessed file to matz@suse.de,
which exhibits such behaviour (and if you even have some dumps from the
profiler, all the better ;) ).  I can't promise to make it really faster
though (except switching off such obvious time takers like
recolor_spills()), as some of the infrastructure is still missing even in
the regalloc branch.


Ciao,
Michael.