Converted gcc ASMMultU64ByU32DivByU32 inline code to nasm in order to speed up the MSC builds (yes, it makes a noticable difference for TSC).