TorqueGameEngines/Torque3D
1
0
Fork 0
mirror of https://github.com/TorqueGameEngines/Torque3D.git synced 2026-03-20 12:50:57 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Test simd math functions for float4 values

Browse source 
beginning implementation of float4 simd functions for x64 and neon
This commit is contained in:
marauder2k7 2026-02-25 15:06:04 +00:00
parent a7d92c344d
commit 6406ca1832
8 changed files with 305 additions and 389 deletions
Download patch file Download diff file Expand all files Collapse all files

33
Engine/source/math/mMathFn.h
View file

@ -37,6 +37,39 @@
extern void MathConsoleInit();
//------------------------------------------------------------------------------
// Runtime ISA detection enum for x86
enum class ISA {
NONE = 0,
SSE2,
SSE41,
AVX,
AVX2
};
// Global variable storing detected ISA
extern ISA gISA;
//------------------------------------------------------------------------------
// Forward declare math_backend
namespace math_backend {
namespace float4 {
void add(const float* a, const float* b, float* r);
void sub(const float* a, const float* b, float* r);
void mul(const float* a, const float* b, float* r);
void mul_scalar(const float* a, float s, float* r);
void div(const float* a, const float* b, float* r);
void div_scalar(const float* a, float s, float* r);
float dot(const float* a, const float* b);
void lerp(const float* from, const float* to, float f, float* r);
void addArray(const class Point4F* a, const class Point4F* b, class Point4F* r, size_t n);
void subArray(const class Point4F* a, const class Point4F* b, class Point4F* r, size_t n);
void mulArray(const class Point4F* a, const class Point4F* b, class Point4F* r, size_t n);
void mulArrayScalar(const class Point4F* a, float s, class Point4F* r, size_t n);
void lerpArray(const class Point4F* from, const class Point4F* to, float f, class Point4F* r, size_t n);
}
} // namespace math_backend
//--------------------------------------
// Installable Library Prototypes
extern S32 (*m_mulDivS32)(S32 a, S32 b, S32 c);