[llvm] [AArch64] Disable Pre-RA Scheduler for Neoverse V2 (PR #127784)

[Sjoerd Meijer via llvm-commits]

sjoerdmeijer wrote:

> but there are times when scheduling on OoO can be helpful 

But we don't have the evidence for this. At least, not in the workloads that we see. Now, we might find one such a case, but is that then representative? The whole hypothesis is that scheduling isn't a win overall and is just eating up compile time (this idea is shared by gcc folks, fwiw).

>Disabling enableMachineScheduler() starts to enable some scheduling out of SDAG, and that might be hitting crashed now. 

I also don't know what this crash is.

> There is a chance this works differently to specifying the option, can you give it another test?

Sorry, the experiment is unclear to me, what is it exactly?

> Is it possible to add a test case for the problem you saw where no scheduling was better than with?

I am afraid it will be difficult to reduce and come up with something small, but can give it a try.

https://github.com/llvm/llvm-project/pull/127784