[llvm] [PassBuilder] Remove LoopFullUnrollPass from before LoopVectorize in the FullLTO pipeline (PR #197176)

[Momchil Velikov via llvm-commits]

momchil-velikov wrote:

> > > Can you get some benchmark numbers on this for the Graviton4? Just to check there's no really big regressions. I've kicked off a run on some RISC-V hardware, I should have results tomorrow
> > 
> > 
> > I don't have benchmarks on individual patches, just on everything together, results [here](https://github.com/llvm/llvm-project/pull/195886#issuecomment-4830998464)
> 
> Is it possible to get a benchmark run on this individual patch? IIUC this patch needs to land first before #195886 so we need to make sure we don't regress in the meantime

Yeah, it's not a problem technically (I'll make a run and post the results later), but really these patches are not supposed to exist individually.

https://github.com/llvm/llvm-project/pull/197176