[llvm] [AMDGPU][SDAG] Support source modifiers as integer on select (PR #147325)

[Matt Arsenault via llvm-commits]

https://github.com/arsenm commented:

There are 2 ways to go about this. Option 1 is what you've done here, where you hack up the integer operations to use the FP types. Option 2 would be to start directly selecting the modifiers through the integer patterns. There are plusses and minuses to each, and the main deciding factor would be how this interferes with other combines.

The examples here don't have any other context to make a judgement on. We also know have SimplifyDemandedBits through the FP sign bit operators, so maybe it matters less now.



https://github.com/llvm/llvm-project/pull/147325