[llvm] [RISCV64] liveness analysis (PR #167454)

[via llvm-commits]

hiraditya wrote:

> What is your expectation for the impact because I've tried this with SPEC and I've not succeeded in getting any meaningful difference in dynamic instruction count for SPEC2017 `-march=rva22u64 -O3` (i.e. the diff rounds to 0.00%). All below done with the [2e0659b](https://github.com/llvm/llvm-project/commit/2e0659bfe0ff506a5ece9f2b3fc1ee397bef3345) version of the PR.

Thanks for running the numbers. I'll need to investigate which passes can benefit from precise liveness and hopefully we get some benefits.

> 
> Also, please remember to give guidance on what combinations of flags you expect people to use in order to test this properly! My attempts:

Good point. I've updated the defaults so update liveness (local+global) all the time and disable verifier.


https://github.com/llvm/llvm-project/pull/167454