`-mavx` should imply `-msse4.2` (or, implications between x86 CPU feature flags)

Summary

If a program is compiled with -mavx, the CPU running it should also support SSE4.2, thus GHC can emit popcnt instruction.

More generally, GHC should have a better knowledge of implications between CPU features. As a note, LLVM's knowledge is written at: https://github.com/llvm/llvm-project/blob/main/llvm/lib/Target/X86/X86.td

Maybe SseVersion in compiler/GHC/Platform.hs should include AVX family.

Steps to reproduce

$ cat popcounttest.hs
import Data.Bits

{-# NOINLINE foo #-}
foo :: Int -> Int
foo x = 1 + popCount x

main = print (foo 42)
$ ghc -S -O2 -fforce-recomp popcounttest.hs
$ grep -i popcnt popcounttest.s
        call hs_popcnt64
$ ghc -S -O2 -fforce-recomp -msse4.2 popcounttest.hs
$ grep -i popcnt popcounttest.s
        popcnt %r14,%rax
$ ghc -S -O2 -fforce-recomp -mavx popcounttest.hs
$ grep -i popcnt popcounttest.s
        call hs_popcnt64

Expected behavior

GHC should emit popcnt when compiling with -mavx.

Environment

• GHC version used: 9.10.1 and 9.11.20240614 (ce76bf78)
• Operating System: Ubuntu 22.04
• System Architecture: x86_64