HH-Math provides a set of fast transcendental functions, SIMD-optimized for each platform.
There are versions for:
SIMD:
calar
SIMD::Vector
SIMD:
treams
If possible, prefer using the streams versions. They operate on arrays of values, and are usually faster than the Vector or Scalar forms. If not, prefer using the SIMD::Vector forms.
These functions can be found in the FullPrecision and LowPrecision namespaces inside the SIMD namespace.
For example, to call the FullPrecision Exp2 function on a vector input, just call:
SIMD::Float4 v = SIMD::Float4::LoadAligned16(srcPtr);
SIMD::Float4 result = SIMD::FullPrecision::Exp2(v);
The FullPrecision functions generally have good relative precision, and are still fast.
The LowPrecision functions have lower precision, but are often much faster.
In general, both versions outperform the native math library implementations (especially on consoles, Even for the XMath implementations on Xbox360).
See the tables below for more details. You will also find more precise per-platform timings in each of the function's pages.
The following functions are available:
calar 
