[llvm] [DAG] foldAddToAvg - split nested m_Reassociatable matchers (PR #171681)

[Min-Yih Hsu via llvm-commits]

mshockwave wrote:

> What's interesting is that this has had a negligible effect on the compile time tracker numbers

Interesting, perhaps changing the container type (from `SmallVector<SDValue>`) as you pointed out in another place has a larger impact, which is also related to the report compiler crash regarding overflowing SmallVector's stack. 

https://github.com/llvm/llvm-project/pull/171681