llvm-project/bolt/lib
spupyrev 4609f60ebc [BOLT] Avoid pointless loop rotation
It seems the earlier implementation does not follow the description
in LoopRotationPass.h: It rotates loops even if they are already laid out
correctly. The diff adjusts the behaviour.

Given that the impact of LoopInversionPass is minor, this change won't
yield significant perf differences. Tested on clang-10: there seems to be a
0.1%-0.3% cpu win and a small reduction of branch misses.

**Before:**
BOLT-INFO: 120 Functions were reordered by LoopInversionPass

**After:**
BOLT-INFO: 79 Functions were reordered by LoopInversionPass

Reviewed By: yota9

Differential Revision: https://reviews.llvm.org/D121921
2022-03-22 12:42:42 -07:00
..
Core [BOLT] Set cold sections alignment explicitly 2022-03-15 22:12:17 +03:00
Passes [BOLT] Avoid pointless loop rotation 2022-03-22 12:42:42 -07:00
Profile Cleanup LLVMDWARFDebugInfo 2022-02-15 09:16:03 +01:00
Rewrite [BOLT] LongJmp: Fix hot text section alignment 2022-03-16 15:57:46 +03:00
RuntimeLibs [BOLT][NFC] Fix braces usage in the rest of the codebase 2021-12-28 18:43:53 -08:00
Target [BOLT][NFC] Use X86 mnemonic tables 2022-03-18 01:52:11 -07:00
Utils [BOLT] Set cold sections alignment explicitly 2022-03-15 22:12:17 +03:00
CMakeLists.txt Rebase: [NFC] Refactor sources to be buildable in shared mode 2021-10-08 11:47:10 -07:00