Actually, I looked at doing just that. And while this is possible, my experience suggest that gcc's built-ins don't do *quite* the same thing. For example, __builtin_ffs() returns 0 if no bits are set, or the 1-based bit position if a bit is set. BitScanForward return a BOOL to indicate no bits are set, and a 0-based bit position is returned in the pointer passed as p1.
Yes, you can write a routine to jigger the output of the __builtin to match the behavior MSVC programmers expect, but the code tends to be sub-optimal. For example (IIRC), when you subtract 1 from ffs to turn it back to 0-based, the asm that ultimately gets produced uses the BSF instruction, adds one, then subtracts 1.<sigh> If I were talking with the MS team when they were first writing BitScanForward, I would have encouraged them to go the way ffs does. However, if the goal here is to support MSVC ports, I think the inlines are a better (more efficient) way to go. dw On 3/28/2013 3:52 AM, Václav Zeman wrote: > Why are those intrinsics even implemented using inline assembler > instead of GCC's own built-in functions like __builtin_ffs() for > _BitScanForward(), __sync_add_and_fetch() for _InterlockedAdd(), etc.? > ------------------------------------------------------------------------------ Own the Future-Intel(R) Level Up Game Demo Contest 2013 Rise to greatness in Intel's independent game demo contest. Compete for recognition, cash, and the chance to get your game on Steam. $5K grand prize plus 10 genre and skill prizes. Submit your demo by 6/6/13. http://altfarm.mediaplex.com/ad/ck/12124-176961-30367-2 _______________________________________________ Mingw-w64-public mailing list Mingw-w64-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
