Recommit: "[MLIR][NVVM] Replace fdiv on fp16 with promoted (fp32)...

Recommit: "[MLIR][NVVM] Replace fdiv on fp16 with promoted (fp32) multiplication with reciprocal plus one (conditional) Newton iteration."

This change rolls bcfc0a90 forward (i.e., reverting 369ce54b) with fixed CMakeLists.txt.