TorqueGameEngines/Torque3D
1
0
Fork 0
mirror of https://github.com/TorqueGameEngines/Torque3D.git synced 2026-03-19 12:20:57 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

ISA backends float3 and float4 - cleanup history squash

Browse source 
working for both neon32 and neon64

Update math_backend.cpp

further sse simd additions

avx2 float3 added
added normalize_magnitude
added divide fast to float3 may copy to float4

move static spheremesh to drawSphere (initialize on first use) so platform has a chance to load the math backend

all float3 and float4 functions and isas

completed all options of float3 and float4 functions in isas and math_c
neon still to be done but that will be on mac.

Update math_backend.cpp

mac isa neon update

added float3
restructured the classes to look more like the final version of the x86 classes

linux required changes

Update build-macos-clang.yml

Update build-macos-clang.yml

Revert "Update build-macos-clang.yml"

This reverts commit 29dfc567f4.

Revert "Update build-macos-clang.yml"

This reverts commit 2abad2b4ca.

Update CMakeLists.txt

fix macs stupid build

remove god awful rolling average from frame time tracker....

use intrinsic headers instead

each isa implementation now uses a header for that isa's intrinsic functions these are then used in the impl files. This will make it easier for matrix functions when those are implemented.

fixed comment saying 256 when it should be 512 for avx512

consolidated initializers for function tables

Update neon_intrinsics.h

fixes for some neon intrinsics no idea if this is the best way to do these but they work at least

v_cross is especially messy at the moment we basically just do it as a c math function need to look into getting this done correctly
This commit is contained in:
marauder2k7 2026-02-26 16:45:13 +00:00
parent 73ed502ac9
commit 67f12311d4
36 changed files with 1481 additions and 419 deletions
Download patch file Download diff file Expand all files Collapse all files

50
Engine/source/math/isa/sse41/float4.cpp
View file

@ -1,49 +1,5 @@
#include "sse41_intrinsics.h"
#include "float4_dispatch.h"
#include <smmintrin.h> // SSE41 intrinsics
namespace
{
typedef __m128 f32x4;
// Load 4 floats from memory into a SIMD register
inline f32x4 v_load(const float* p) { return _mm_loadu_ps(p); }
// Store 4 floats from SIMD register back to memory
inline void v_store(float* dst, f32x4 v) { _mm_storeu_ps(dst, v); }
// Broadcast a single float across all 4 lanes
inline f32x4 v_set1(float s) { return _mm_set1_ps(s); }
// Element-wise multiply
inline f32x4 v_mul(f32x4 a, f32x4 b) { return _mm_mul_ps(a, b); }
// Element-wise divide
inline f32x4 v_div(f32x4 a, f32x4 b) { return _mm_div_ps(a, b); }
// Element-wise add
inline f32x4 v_add(f32x4 a, f32x4 b) { return _mm_add_ps(a, b); }
// Element-wise subtract
inline f32x4 v_sub(f32x4 a, f32x4 b) { return _mm_sub_ps(a, b); }
// Horizontal sum of all 4 elements (for dot product, length, etc.)
inline float v_hadd4(f32x4 a)
{
__m128 t1 = _mm_hadd_ps(a, a); // sums pairs: [a0+a1, a2+a3, ...]
__m128 t2 = _mm_hadd_ps(t1, t1); // sums again: first element = a0+a1+a2+a3
return _mm_cvtss_f32(t2); // extract first element
}
// specialized dot product for SSE4.1
float float4_dot_sse41(const float* a, const float* b)
{
f32x4 va = _mm_loadu_ps(a);
f32x4 vb = _mm_loadu_ps(b);
__m128 dp = _mm_dp_ps(va, vb, 0xF1); // multiply all 4, sum all 4, lowest lane
return _mm_cvtss_f32(dp);
}
}
#include "float4_impl.inl"
@ -58,10 +14,12 @@ namespace math_backend::float4::dispatch
gFloat4.mul_scalar = float4_mul_scalar_impl;
gFloat4.div = float4_div_impl;
gFloat4.div_scalar = float4_div_scalar_impl;
gFloat4.dot = float4_dot_sse41;
gFloat4.dot = float4_dot_impl;
gFloat4.length = float4_length_impl;
gFloat4.lengthSquared = float4_length_squared_impl;
gFloat4.normalize = float4_normalize_impl;
gFloat4.normalize_mag = float4_normalize_mag_impl;
gFloat4.lerp = float4_lerp_impl;
gFloat4.cross = float4_cross_impl;
}
}