This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tree-ssa] POOMA compile time / memory requirement comparison

Steven Bosscher wrote:
On Wednesday 05 May 2004 20:41, Richard Guenther wrote:

Using the suggested -fno-gcse -fno-cse-follow-jumps -fno-cse-skip-blocks
-fno-web, -O2 compilation times improve about 8%.  This leaves us with



Why not -fno-delete-null-pointer-checks?  And -fweb is enabled on mainline
and helps tree-ssa iirc, so that should really stay enabled...

-fno-delete-null-pointer-checks doesn't help. It seems to be cheap.

I thought we have life-splitting enabled in the expander now - so web seemed unnecessary.

10% to improve elsewhere (I guess, f.i. preserving the CFG could help,
as well as removing some first invocations of CSE).



As well as disabling some stuff hidden behind -fexpensive-optimizations,
and investigating if we really need to call delete_trivially_dead_insns
so many times (and iteratively, see cfgcleanup!).  And some other stuff
like addressof, as discussed here some time ago.  There are still a lot
of things we have to try...

Yes. I'm confident we're going to gain more than the current 20% regression. It just needs people remind others about the regressions.

Richard.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]