alright I found the bugs in GPU3D.cpp and GPU.cpp
GPU3D.cpp should read like this at line 695
#ifdef NEONSOFTGPU_ENABLED
void MatrixMult4x4(s32* m, s32* s);
void MatrixMult4x3(s32* m, s32* s);
void MatrixMult3x3(s32* m, s32* s);
void MatrixScale(s32* m, s32* s);
void MatrixTranslate(s32* m...