summaryrefslogtreecommitdiffstats
path: root/src/math/math.cpp
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/math/math.cpp128
1 files changed, 128 insertions, 0 deletions
diff --git a/src/math/math.cpp b/src/math/math.cpp
index 0cfc2ce9..4792fdd0 100644
--- a/src/math/math.cpp
+++ b/src/math/math.cpp
@@ -1,9 +1,124 @@
#include "common.h"
#include "Quaternion.h"
+#include "VuVector.h"
// TODO: move more stuff into here
+
+void TransformPoint(CVuVector &out, const CMatrix &mat, const CVuVector &in)
+{
+#ifdef GTA_PS2
+ __asm__ __volatile__("\n\
+ lqc2 vf01,0x0(%2)\n\
+ lqc2 vf02,0x0(%1)\n\
+ lqc2 vf03,0x10(%1)\n\
+ lqc2 vf04,0x20(%1)\n\
+ lqc2 vf05,0x30(%1)\n\
+ vmulax.xyz ACC, vf02,vf01\n\
+ vmadday.xyz ACC, vf03,vf01\n\
+ vmaddaz.xyz ACC, vf04,vf01\n\
+ vmaddw.xyz vf06,vf05,vf00\n\
+ sqc2 vf06,0x0(%0)\n\
+ ": : "r" (&out) , "r" (&mat) ,"r" (&in): "memory");
+#else
+ out = mat * in;
+#endif
+}
+
+void TransformPoint(CVuVector &out, const CMatrix &mat, const RwV3d &in)
+{
+#ifdef GTA_PS2
+ __asm__ __volatile__("\n\
+ ldr $8,0x0(%2)\n\
+ ldl $8,0x7(%2)\n\
+ lw $9,0x8(%2)\n\
+ pcpyld $10,$9,$8\n\
+ qmtc2 $10,vf01\n\
+ lqc2 vf02,0x0(%1)\n\
+ lqc2 vf03,0x10(%1)\n\
+ lqc2 vf04,0x20(%1)\n\
+ lqc2 vf05,0x30(%1)\n\
+ vmulax.xyz ACC, vf02,vf01\n\
+ vmadday.xyz ACC, vf03,vf01\n\
+ vmaddaz.xyz ACC, vf04,vf01\n\
+ vmaddw.xyz vf06,vf05,vf00\n\
+ sqc2 vf06,0x0(%0)\n\
+ ": : "r" (&out) , "r" (&mat) ,"r" (&in): "memory");
+#else
+ out = mat * in;
+#endif
+}
+
+void TransformPoints(CVuVector *out, int n, const CMatrix &mat, const RwV3d *in, int stride)
+{
+#ifdef GTA_PS3
+ __asm__ __volatile__("\n\
+ paddub $3,%4,$0\n\
+ lqc2 vf02,0x0(%2)\n\
+ lqc2 vf03,0x10(%2)\n\
+ lqc2 vf04,0x20(%2)\n\
+ lqc2 vf05,0x30(%2)\n\
+ ldr $8,0x0(%3)\n\
+ ldl $8,0x7(%3)\n\
+ lw $9,0x8(%3)\n\
+ pcpyld $10,$9,$8\n\
+ qmtc2 $10,vf01\n\
+ 1: vmulax.xyz ACC, vf02,vf01\n\
+ vmadday.xyz ACC, vf03,vf01\n\
+ vmaddaz.xyz ACC, vf04,vf01\n\
+ vmaddw.xyz vf06,vf05,vf00\n\
+ add %3,%3,$3\n\
+ ldr $8,0x0(%3)\n\
+ ldl $8,0x7(%3)\n\
+ lw $9,0x8(%3)\n\
+ pcpyld $10,$9,$8\n\
+ qmtc2 $10,vf01\n\
+ addi %1,%1,-1\n\
+ addiu %0,%0,0x10\n\
+ sqc2 vf06,-0x10(%0)\n\
+ bnez %1,1b\n\
+ ": : "r" (out) , "r" (n), "r" (&mat), "r" (in), "r" (stride): "memory");
+#else
+ while(n--){
+ *out = mat * *in;
+ in = (RwV3d*)((uint8*)in + stride);
+ out++;
+ }
+#endif
+}
+
+void TransformPoints(CVuVector *out, int n, const CMatrix &mat, const CVuVector *in)
+{
+#ifdef GTA_PS2
+ __asm__ __volatile__("\n\
+ lqc2 vf02,0x0(%2)\n\
+ lqc2 vf03,0x10(%2)\n\
+ lqc2 vf04,0x20(%2)\n\
+ lqc2 vf05,0x30(%2)\n\
+ lqc2 vf01,0x0(%3)\n\
+ nop\n\
+ 1: vmulax.xyz ACC, vf02,vf01\n\
+ vmadday.xyz ACC, vf03,vf01\n\
+ vmaddaz.xyz ACC, vf04,vf01\n\
+ vmaddw.xyz vf06,vf05,vf00\n\
+ lqc2 vf01,0x10(%3)\n\
+ addiu %3,%3,0x10\n\
+ addi %1,%1,-1\n\
+ addiu %0,%0,0x10\n\
+ sqc2 vf06,-0x10(%0)\n\
+ bnez %1,1b\n\
+ ": : "r" (out) , "r" (n), "r" (&mat) ,"r" (in): "memory");
+#else
+ while(n--){
+ *out = mat * *in;
+ in++;
+ out++;
+ }
+#endif
+}
+
+
void
CVector2D::Normalise(void)
{
@@ -87,6 +202,7 @@ CMatrix::Reorthogonalise(void)
CMatrix&
Invert(const CMatrix &src, CMatrix &dst)
{
+ // TODO: VU0 code
// GTA handles this as a raw 4x4 orthonormal matrix
// and trashes the RW flags, let's not do that
// actual copy of librw code:
@@ -117,6 +233,7 @@ Invert(const CMatrix &src, CMatrix &dst)
CVector
operator*(const CMatrix &mat, const CVector &vec)
{
+ // TODO: VU0 code
return CVector(
mat.m_matrix.right.x * vec.x + mat.m_matrix.up.x * vec.y + mat.m_matrix.at.x * vec.z + mat.m_matrix.pos.x,
mat.m_matrix.right.y * vec.x + mat.m_matrix.up.y * vec.y + mat.m_matrix.at.y * vec.z + mat.m_matrix.pos.y,
@@ -126,6 +243,7 @@ operator*(const CMatrix &mat, const CVector &vec)
CMatrix
operator*(const CMatrix &m1, const CMatrix &m2)
{
+ // TODO: VU0 code
CMatrix out;
RwMatrix *dst = &out.m_matrix;
const RwMatrix *src1 = &m1.m_matrix;
@@ -145,9 +263,18 @@ operator*(const CMatrix &m1, const CMatrix &m2)
return out;
}
+CMatrix&
+CMatrix::operator*=(CMatrix const &rhs)
+{
+ // TODO: VU0 code
+ *this = *this * rhs;
+ return *this;
+}
+
const CVector
Multiply3x3(const CMatrix &mat, const CVector &vec)
{
+ // TODO: VU0 code
return CVector(
mat.m_matrix.right.x * vec.x + mat.m_matrix.up.x * vec.y + mat.m_matrix.at.x * vec.z,
mat.m_matrix.right.y * vec.x + mat.m_matrix.up.y * vec.y + mat.m_matrix.at.y * vec.z,
@@ -179,6 +306,7 @@ CQuaternion::Slerp(const CQuaternion &q1, const CQuaternion &q2, float theta, fl
w1 = Sin((1.0f - t) * theta) * invSin;
w2 = Sin(t * theta) * invSin;
}
+ // TODO: VU0 code
*this = w1*q1 + w2*q2;
}
}