Optimize various masked vector operations for AVX512 target.

- In continuation to Initial X86 backend support for masked vector operations (JDK-8262356), extend backed to support following masked operations.
   - Vector Rotate left/right.
   - Vector Left/Right Arithmetic/Logical shift with immediate shift count.
   - Vector SquareRoot operation over Float/Double primitive types.
   - Optimize VectorMask.not operation.
 - Enable masking optimizations for X86 KNL target.