[llvm] [AMDGPU] fix SIPeepholeSDWA optimization for fp16 (PR #109395)

[Matt Arsenault via llvm-commits]

arsenm wrote:

Looks equivalent to me. The BFE is extracting 8 bits at offset 0, and sign extending it. That should be the same thing as sign extending BYTE_0. 

https://github.com/llvm/llvm-project/pull/109395