[llvm] [AArch64][machine-scheduler][Neoverse-N2] fdiv is blocking (PR #119206)

[Simon Wallis via llvm-commits]

simonwallis2 wrote:

Hi Rin, I have tested this patch on the floating point workloads that I have available and I have not seen any perf regressions.
The other kinds of divide - integer, asimd and sve - I have not added to the patch and have not checked for performance regressions on suitable workloads. I am not sure that I have a suitable workload for asimd and sve divides - do you have such a case?
Deferring such changes to a future commit will reduce the chance of this patch being landed then reverted for performance reasons unrelated to the fp change.


https://github.com/llvm/llvm-project/pull/119206