* [QNN] Optimize lowering for requantize and FixedPointMultiply.

* Add check for requantize scale gt 1.

* Added test case.