On Thu, Oct 4, 2012 at 5:30 AM, Vladimir Makarov <vmaka...@redhat.com> wrote: > I was going to look at this code too but I was interesting in generation of > less points and live ranges. It is strange that in my profiles, > remove_some_program_points_and_update_live_ranges takes 0.6% of compiler > time on these huge tests. So I was not interesting to speed up the > function and may be therefore you have no visible change in compilation > time.
Right. The compression algorithm doesn't care much about the initial number of program points, only about the number of live ranges before and after compression. I had expected a bigger effect on the number of live ranges before compression. 0.6% sounds really very different from my timings. How much time does create_start_finish_chains take for you? > I don't object the idea of the patch. I need some time to look at it (the > different results on a function is a bit scary for me) and check simulator > times on other tests. Understood. FWIW you can find the test results from before and after the patch here: http://gcc.gnu.org/ml/gcc-testresults/2012-10/msg00267.html http://gcc.gnu.org/ml/gcc-testresults/2012-10/msg00303.html Ciao! Steven