[llvm] [LLVM][DecoderEmitter] Add option to use lambdas in decodeToMCInst (PR #144814)

[Rahul Joshi via llvm-commits]

jurahul wrote:

Note: Its possible for further refinements here, like using lambda only when the # of switch cases is  > a threshold, and also adding some a benchmark (using AMDGPU as a target) to measure perf delta between switch-case vs lambda version (I am assuming the lambda version is slower), but I am thinking that can come in future based on needs after this basic feature is adopted and we have some more data.

Also nothing that this is code is (AFAIK) not a part of the "regular" compilation flow (i.e., not built into llc).

https://github.com/llvm/llvm-project/pull/144814