[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 Andrew Pinski changed: What|Removed |Added Status|NEW |RESOLVED Known to work||10.1.0 Resolution|--- |FIXED Target Milestone|--- |10.0 --- Comment #9 from Andrew Pinski --- Fixed in GCC 10.1.0. Part of it was fixed with r6-3841-g84ff66b884e9 . The other part which fixed this was improving EVRP.
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 --- Comment #8 from Eric Gallager --- Does the new combine2 pass proposed for GCC 10 fix any of this?
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 --- Comment #7 from Jim Wilson --- I didn't realize it was assigned to me. I changed it to unassigned.
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 --- Comment #6 from Eric Gallager --- (In reply to Jim Wilson from comment #5) > I haven't done any further work on this. The reason I was asking was because this bug shows up in my search of bugs with an assignee but that aren't marked as ASSIGNED. So, should I change the status to ASSIGNED here, or leave it as NEW and remove you as the assignee?
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 --- Comment #5 from Jim Wilson --- I haven't done any further work on this.
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 Eric Gallager changed: What|Removed |Added CC||egallager at gcc dot gnu.org --- Comment #4 from Eric Gallager --- (In reply to Jim Wilson from comment #3) > This ior/shift/xor optimization be done during combine with a simplify-rtx.c > patch. I wrote a prototype and tested it. Combine canonicalizes > shift/logical as logical/shift, so we actually have to look for > ior/xor/shift. I see that the optimization does not happen on arm and > aarch64 because the 0x4002 value does not fit into the immediate range for > logical ops, gets loaded into a register, pulled out of the loop, and hence > is not available to combine. We would have to perform the optimization > earlier for it to work for arm/aarch64. I tried MIPS, and see that because > MIPS promotes HImode to SImode we don't have enough info to prove that the > opt is safe. We need type info to make this work. This prototype does work > for the x86 target which has both 16-bit immediates and HImode instructions. > > Next I wrote a prototype for match-and-simplify. This one works for all 4 > targets. We have to handle ior/shift/xor as match-and-simplify does not > canonicalize logical/shift. I noticed that if I change the xor to an ior, > then it gets optimized in combine because of the canonicalization for mips > and x86, but not for arm/aarch64 again because the constant was pulled out > of the loop. So it seems that match-and-simplify should canonicalize > shift/logical to logical/shift too. That would also reduce the number of > patterns we need to match when performing this ior/shift/xor optimization. > > These are prototype patches that need a bit more work to handle more cases, > and to prove that they test all conditions necessary to make them safe. > This is my first attempt to use the wi:: functions, so there might be a > better way to do this. Have you made further attempts since?
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 Jim Wilson wilson at gcc dot gnu.org changed: What|Removed |Added CC||wilson at gcc dot gnu.org --- Comment #2 from Jim Wilson wilson at gcc dot gnu.org --- Created attachment 35398 -- https://gcc.gnu.org/bugzilla/attachment.cgi?id=35398action=edit prototype patches for ior/shift/xor optimization
[Bug rtl-optimization/26190] combine misses some distributivity
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190 --- Comment #3 from Jim Wilson wilson at gcc dot gnu.org --- This ior/shift/xor optimization be done during combine with a simplify-rtx.c patch. I wrote a prototype and tested it. Combine canonicalizes shift/logical as logical/shift, so we actually have to look for ior/xor/shift. I see that the optimization does not happen on arm and aarch64 because the 0x4002 value does not fit into the immediate range for logical ops, gets loaded into a register, pulled out of the loop, and hence is not available to combine. We would have to perform the optimization earlier for it to work for arm/aarch64. I tried MIPS, and see that because MIPS promotes HImode to SImode we don't have enough info to prove that the opt is safe. We need type info to make this work. This prototype does work for the x86 target which has both 16-bit immediates and HImode instructions. Next I wrote a prototype for match-and-simplify. This one works for all 4 targets. We have to handle ior/shift/xor as match-and-simplify does not canonicalize logical/shift. I noticed that if I change the xor to an ior, then it gets optimized in combine because of the canonicalization for mips and x86, but not for arm/aarch64 again because the constant was pulled out of the loop. So it seems that match-and-simplify should canonicalize shift/logical to logical/shift too. That would also reduce the number of patterns we need to match when performing this ior/shift/xor optimization. These are prototype patches that need a bit more work to handle more cases, and to prove that they test all conditions necessary to make them safe. This is my first attempt to use the wi:: functions, so there might be a better way to do this.
[Bug rtl-optimization/26190] combine misses some distributivity
--- Comment #1 from pinskia at gcc dot gnu dot org 2006-02-25 19:40 --- Confirmed. -- pinskia at gcc dot gnu dot org changed: What|Removed |Added Status|UNCONFIRMED |NEW Ever Confirmed|0 |1 Last reconfirmed|-00-00 00:00:00 |2006-02-25 19:40:32 date|| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190
[Bug rtl-optimization/26190] combine misses some distributivity
-- bonzini at gnu dot org changed: What|Removed |Added BugsThisDependsOn||16798, 18395 Severity|normal |enhancement Priority|P3 |P1 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190
[Bug rtl-optimization/26190] combine misses some distributivity
-- pinskia at gcc dot gnu dot org changed: What|Removed |Added Keywords||missed-optimization Priority|P1 |P5 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26190
