I’m wondering if there really is no 128-bit division intrinsic function in Visual C++?
There is a 64×64=128 bit multiplication intrinsic function called _umul128(), which nicely matches the MUL x64 assembler instruction.
Naturally, I assumed there would be a 128/64=64 bit division intrinsic as well (modelling the DIV instruction), but to my amazement neither Visual C++ nor Intel C++ seem to have it, at least it’s not listed in intrin.h.
Can someone confirm that? I tried grep’ing for the function names in the compiler executable files, but couldn’t find _umul128 in the first place, so I guess I looked in the wrong spot.
Update: at least I have now found the pattern umul128 (without the leading underscore) in c1.dll of Visual C++ 2010. All the other intrinsics are listed around it, but unfortunately no “udiv128” or the like 🙁 So it seems they really have “forgotten” to implement it.
To clarify: I’m not only looking for a 128-bit data type, but a way to divide a 128-bit scalar int by a 64-bit int in C++. Either an intrinsic function or native 128-bit integer support would solve my problem.
Edit: The answer is no, there is no _udiv128 intrinsic in Visual Studio 2010 up to 2017, but it is available in Visual Studio 2019 RTM
I am no expert, but I dug this up:
http://research.swtch.com/2008/01/division-via-multiplication.html
Interesting stuff. Hope it helps.
EDIT: This is insightful too: http://www.gamedev.net/topic/508197-x64-div-intrinsic/