https://bugs.llvm.org/show_bug.cgi?id=51857
Bug ID: 51857
Summary: Interpreter project is up to 15% slower with clang 13
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: Scalar Optimizations
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
I maintain a relatively simple interpreter project (single main function with
big switch statement).
I tested clang 13 RC and it gives 0-15% slowdown depending on the workload. In
none of the cases the clang 13 is faster than clang 12.
I tried to look at generated assembly and LLVM IR but I cannot stop anything
interesting there. Maybe just that clang 13 produces smaller code probably due
to better register allocation / spilling.
Also perf tool does not show anything specific.
I'm happy to provide additional data if you can tell we where to start and what
would be useful.
I also started bisecting LLVM from 12.x to 13.x. So far I have found single
commit that decreases performance by ~6%:
[LVI] Don't bail on overdefined value in select
https://github.com/llvm/llvm-project/commit/72e0846ef87d0d3b5960238bc47fc0cc6f04d848
https://reviews.llvm.org/rG72e0846ef87d0d3b5960238bc47fc0cc6f04d848
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
