[llvm] [AArch64] Neoverse V2 FeatureDisableLatencySchedHeuristic (PR #140897)

[Igor Kirillov via llvm-commits]

igogo-x86 wrote:

LLVM IR is exactly the same before the SLP patch (#136329) and after my fix (#140897). However, the performance was still worse. So I looked at the assembly and saw that the instructions were identical, but the order was different. I then bisected the issue down to your change.

Here are the `perf` sample data:

| Function                       | Old Good Times | After SLP Patch | My Fix with Your Patch | My Fix without Your Patch |
|-------------------------------|----------------|------------------|----------------------|----------------------|
| `sym.x264_pixel_satd_16x16`   | 11488          | 12236            | 12044                | 11582                |
| `sym.x264_pixel_satd_8x8`     | 7192           | 7884             | 6949                 | 6644                 |

P.S. Something improved `sym.x264_pixel_satd_8x8` in the meantime.

https://github.com/llvm/llvm-project/pull/140897