[llvm] [AArch64] Update latencies for Cortex-A510 scheduling model (PR #87293)

[David Green via llvm-commits]

https://github.com/davemgreen commented:

Thanks for the changes. I believe that there are implementations of both Cortex-A510 r0 and Cortex-A510 r1, but I think it makes sense to use the newer numbers going forward. I'm not sure which of them are documentation fixes and which were real improvements, but I checked through the numbers and most of them looked good.

https://github.com/llvm/llvm-project/pull/87293