I am interested in bitwise operations. I use ^ & | << >> ~ the most. And + - from time to time. Never / * %.

What about firstbitlow/high?

You can get some of that information here for NVIDIA:

https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#arithmetic-instructions

For AMD GCN parts, all the operators that you said you use are full-speed. 32-bit multiplication is 4x slower than full-rate. / and % are microcoded operations that take much longer (~15-20x) slower than full-rate (unless the compiler happens to know the denominator and can optimise for it specifically).