[llvm] [AMDGPU] Replace zext pattern from reg_sequence to cvt_u32_u16 (PR #208045)

[Guo Chen via llvm-commits]

broxigarchen wrote:

> LGTM on the grounds it is a temporary workaround for a compile time issue and we will reintroduce the patterns when the root cause can be fixed

Thanks! For this patch I would consider it not a workaround though since using cvt_u32_u16 provide better code quality. Marked a few in the lit tests

https://github.com/llvm/llvm-project/pull/208045