c++ - Avoiding DirectXMath XMStore/load -
c++ - Avoiding DirectXMath XMStore/load -
i've done opengl , learning directx11. 1 of things in new math library presence of load/store methods vectors , matrices (for illustration http://msdn.microsoft.com/en-us/library/ee415635(v=vs.85).aspx). me quirks me alot having store/load each vector , matrix when want utilize them - alot of bulky code - , wondering if there nice, clean alternatives load/store procedures?
you @ simplemath if helps. it's part of directxtk.
from linked site:
why wrap directxmath? directxmath provides highly optimized vector , matrix math functions, take advantage of sse simd intrinsics when compiled x86/x64, or arm neon instruction set when compiled arm platform such windows rt or windows phone. downside of beingness designed efficient simd usage directxmath can complicated work with. developers must aware of right type usage (understanding difference between simd register types such xmvector vs. memory storage types such xmfloat4), must take care maintain right alignment simd heap allocations, , must construction code avoid accessing individual components simd register. complexity necessary optimal simd performance, want stuff working without much hassle!
you should utilize simplemath if are:
looking c++ math library similar api c# xna types porting existing xna code c# c++ wanting optimize programmer efficiency (simplicity, readability, development speed) @ expense of runtime efficiencyyou should go straight underlying directxmath api if you:
want create fastest possible code enjoy lateral thinking needed express algorithms in raw simd c++ directx directx-11 directxmath
Comments
Post a Comment