Optimize ulong/double scalar conversion on xarch with AVX512VL instructions.

[kote01-coder]

With some of the new AVX512VL instruction, we can optimize uint32/uint64 to/from single/double conversions within the JIT. Changes can be made in genIntToFloatCast and genFloatToIntCast in codegenxarch.cpp. Currently, xarch requires more complex instruction sequences that can be replaced with the following instructions when AVX512VL is available:

vcvtusi2sd for uint32/uint64 => double conversion
vcvtusi2ss for uint32/uint64 => single conversion
vcvtsd2usi for double => uint32/uint64 conversion
vcvtss2usi for single => uint32/uint64 conversion