A "simple" math library for rendering operations - "simple" for cross-platform speed.
Vectors, matrices, and quaternions. For now.
It has many functions, but I'll definitely be adding MANY more. Check include/library/sml.h.
Lots of union-abuse to support syntax like GLSL's "swizzles".
Code is as simple as possible for API granularity reasons as well as "general" platform-specific optimization reasons (may or may not use platform-specific SIMD intrinsics till I have the time to).
For now, I really do rely on "simple code" and compiler auto-vectorization.
All functions take memory addresses to the data structure[s] to perform an operation on, and a destination. Lots of const-correctness on those. Hopefully the compiler optimizes those, but I cannot give guarantees.
Should also provide batch-operation functions!... I do not have those right now!!!
Library does not handle allocation - might offer allocation pools some day, use IDs instead of pointers, maybe have more speed that way. That day iiis... not today!