On Mon, Dec 7, 2020 at 3:53 PM Eric Botcazou <botca...@adacore.com> wrote: > > Hi, > > the very recent addition of the if_to_switch pass has partially disabled the > optimization I added back in June: > > 2020-06-26 Eric Botcazou <ebotca...@adacore.com> > > * tree-ssa-reassoc.c (dump_range_entry): New function. > (debug_range_entry): New debug function. > (update_range_test): Invoke dump_range_entry for dumping. > (optimize_range_tests_to_bit_test): Merge the entry test in the > bit test when possible and lower the profitability threshold. > > as witnessed by the 3 new failures in the gnat.dg testsuite. It turns out > that both tree-ssa-reassoc.c and tree-switch-conversion.c can turn things into > bit tests so the optimization must also be added to bit_test_cluster::emit. > > The patch also contains a secondary optimization, whereby the full bit-test > sequence is sent to the folder before being gimplified in case there is only > one test, so that the most optimized sequence (bt + jc on x86) can be emitted > like with optimize_range_tests_to_bit_test. > > Bootstrapped/regtested on x86-64/Linux, OK for the mainline?
OK. Richard. > > 2020-12-07 Eric Botcazou <ebotca...@adacore.com> > > PR tree-optimization/96344 > * tree-switch-conversion.c (bit_test_cluster::emit): Compute the > range only if an entry test is necessary. Merge the entry test in > the bit test when possible. Use PREC local variable consistently. > When there is only one test, do a single gimplification at the end. > > -- > Eric Botcazou
