RuiFPB/tyra-linux
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases Wiki Activity

gcc porting - successful compilation

Browse Source
This commit is contained in:
h4570
2021-01-12 19:34:30 +01:00
parent 67bbe2e674
commit bc41afdf89
12 changed files with 249 additions and 254 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/engine/models/math/matrix.cpp
+95 -95
View File
@@ -58,16 +58,16 @@ Vector3 Matrix::operator*(const Vector3 &v) const
float a[4] = {v.x, v.y, v.z, 1.0F};
float res[4];
asm volatile(
"lqc2 vf4, 0x00(%1) \n\t"
"lqc2 vf5, 0x10(%1) \n\t"
"lqc2 vf6, 0x20(%1) \n\t"
"lqc2 vf7, 0x30(%1) \n\t"
"lqc2 vf8, 0x00(%2) \n\t"
"vmulax.xyzw ACC, vf4, vf8 \n\t"
"vmadday.xyzw ACC, vf5, vf8 \n\t"
"vmaddaz.xyzw ACC, vf6, vf8 \n\t"
"vmaddw.xyzw vf9, vf7, vf8 \n\t"
"sqc2 vf9, 0x00(%0) \n\t"
"lqc2 $vf4, 0x00(%1) \n\t"
"lqc2 $vf5, 0x10(%1) \n\t"
"lqc2 $vf6, 0x20(%1) \n\t"
"lqc2 $vf7, 0x30(%1) \n\t"
"lqc2 $vf8, 0x00(%2) \n\t"
"vmulax.xyzw $ACC, $vf4, $vf8 \n\t"
"vmadday.xyzw $ACC, $vf5, $vf8 \n\t"
"vmaddaz.xyzw $ACC, $vf6, $vf8 \n\t"
"vmaddw.xyzw $vf9, $vf7, $vf8 \n\t"
"sqc2 $vf9, 0x00(%0) \n\t"
:
: "r"(res), "r"(this->data), "r"(a));
return Vector3(res[0], res[1], res[2]);
@@ -80,15 +80,15 @@ Vector3 Matrix::operator*(const Vector3 &v) const
void Matrix::identity()
{
asm volatile(
"vsub.xyzw vf4, vf0, vf0 \n\t"
"vadd.w vf4, vf4, vf0 \n\t"
"vmr32.xyzw vf5, vf4 \n\t"
"vmr32.xyzw vf6, vf5 \n\t"
"vmr32.xyzw vf7, vf6 \n\t"
"sqc2 vf4, 0x30(%0) \n\t"
"sqc2 vf5, 0x20(%0) \n\t"
"sqc2 vf6, 0x10(%0) \n\t"
"sqc2 vf7, 0x0(%0) \n\t"
"vsub.xyzw $vf4, $vf0, $vf0 \n\t"
"vadd.w $vf4, $vf4, $vf0 \n\t"
"vmr32.xyzw $vf5, $vf4 \n\t"
"vmr32.xyzw $vf6, $vf5 \n\t"
"vmr32.xyzw $vf7, $vf6 \n\t"
"sqc2 $vf4, 0x30(%0) \n\t"
"sqc2 $vf5, 0x20(%0) \n\t"
"sqc2 $vf6, 0x10(%0) \n\t"
"sqc2 $vf7, 0x0(%0) \n\t"
:
: "r"(this->data));
}
@@ -131,17 +131,17 @@ void Matrix::lookAt(const Vector3 &t_position, const Vector3 &t_target)
VECTOR eye = {t_position.x, t_position.y, t_position.z, 1.0F};
VECTOR obj = {t_target.x, t_target.y, t_target.z, 1.0F};
asm volatile(
"lqc2 vf4, 0x00(%2) # eye \n\t"
"lqc2 vf5, 0x00(%3) # obj \n\t"
"vsub.xyz vf7, vf4, vf5 # view_vec = vf7 \n\t"
"vmove.xyzw vf6, vf0 \n\t"
"vaddw.y vf6, vf0, vf0 # vf6 = { 0.0f, 1.0f, 0.0f, 1.0f } \n\t"
"vopmula.xyz ACC, vf6, vf7 \n\t"
"vopmsub.xyz vf9, vf7, vf6 # vec = vf9 \n\t"
"vopmula.xyz ACC, vf7, vf9 \n\t"
"vopmsub.xyz vf8, vf9, vf7 # up_vec = vf8 \n\t"
"sqc2 vf7, 0x00(%0) # view_vec \n\t"
"sqc2 vf6, 0x00(%1) # up_vec \n\t"
"lqc2 $vf4, 0x00(%2) # eye \n\t"
"lqc2 $vf5, 0x00(%3) # obj \n\t"
"vsub.xyz $vf7, $vf4, $vf5 # view_vec = $vf7 \n\t"
"vmove.xyzw $vf6, $vf0 \n\t"
"vaddw.y $vf6, $vf0, $vf0 # $vf6 = { 0.0f, 1.0f, 0.0f, 1.0f } \n\t"
"vopmula.xyz $ACC, $vf6, $vf7 \n\t"
"vopmsub.xyz $vf9, $vf7, $vf6 # vec = $vf9 \n\t"
"vopmula.xyz $ACC, $vf7, $vf9 \n\t"
"vopmsub.xyz $vf8, $vf9, $vf7 # up_vec = $vf8 \n\t"
"sqc2 $vf7, 0x00(%0) # view_vec \n\t"
"sqc2 $vf6, 0x00(%1) # up_vec \n\t"
:
: "r"(view_vec), "r"(up_vec), "r"(eye), "r"(obj));
Matrix temp;
@@ -270,45 +270,45 @@ void Matrix::setScale(const Vector3 &t_val)
void Matrix::setCamera(const float t_pos[4], const float t_vz[4], const float t_vy[4])
{
// Matrix vf4, vf5, vf6, vf7
// t_pos vf8
// t_vz vf9
// t_vy vf10
// vtmp vf11
// Matrix $vf4, $vf5, $vf6, $vf7
// t_pos $vf8
// t_vz $vf9
// t_vy $vf10
// vtmp $vf11
asm volatile(
"lqc2 vf9, 0x00(%2) \n\t"
"lqc2 vf10, 0x00(%3) \n\t"
"lqc2 $vf9, 0x00(%2) \n\t"
"lqc2 $vf10, 0x00(%3) \n\t"
// mtmp.unit()
"vsub.w vf5, vf0, vf0 # mtmp[1][PW] = 0.0F \n\t"
"vsub.w $vf5, $vf0, $vf0 # mtmp[1][PW] = 0.0F \n\t"
// vtmp.outerProduct(vy, vz);
"vopmula.xyz ACC, vf10, vf9 \n\t"
"vopmsub.xyz vf11, vf9, vf10 \n\t"
"vopmula.xyz $ACC, $vf10, $vf9 \n\t"
"vopmsub.xyz $vf11, $vf9, $vf10 \n\t"
// mtmp[0] = vtmp.normalize();
"vmul.xyz vf12, vf11, vf11 \n\t"
"vaddy.x vf12, vf12, vf12 \n\t"
"vaddz.x vf12, vf12, vf12 \n\t"
"vrsqrt Q, vf0w, vf12x \n\t"
"vsub.xyzw vf4, vf0, vf0 \n\t"
"vmul.xyz $vf12, $vf11, $vf11 \n\t"
"vaddy.x $vf12, $vf12, $vf12 \n\t"
"vaddz.x $vf12, $vf12, $vf12 \n\t"
"vrsqrt $Q, $vf0w, $vf12x \n\t"
"vsub.xyzw $vf4, $vf0, $vf0 \n\t"
"vwaitq \n\t"
"vmulq.xyz vf4, vf11, Q \n\t"
"vmulq.xyz $vf4, $vf11, $Q \n\t"
// mtmp[2] = vz.normalize();
"vmul.xyz vf12, vf9, vf9 \n\t"
"vaddy.x vf12, vf12, vf12 \n\t"
"vaddz.x vf12, vf12, vf12 \n\t"
"vrsqrt Q, vf0w, vf12x \n\t"
"vsub.xyzw vf6, vf0, vf0 \n\t"
"vmul.xyz $vf12, $vf9, $vf9 \n\t"
"vaddy.x $vf12, $vf12, $vf12 \n\t"
"vaddz.x $vf12, $vf12, $vf12 \n\t"
"vrsqrt $Q, $vf0w, $vf12x \n\t"
"vsub.xyzw $vf6, $vf0, $vf0 \n\t"
"vwaitq \n\t"
"vmulq.xyz vf6, vf9, Q \n\t"
"vmulq.xyz $vf6, $vf9, $Q \n\t"
// mtmp[1].outerProduct(mtmp[2], mtmp[0]);
"vopmula.xyz ACC, vf6, vf4 \n\t"
"vopmsub.xyz vf5, vf4, vf6 \n\t"
"vopmula.xyz $ACC, $vf6, $vf4 \n\t"
"vopmsub.xyz $vf5, $vf4, $vf6 \n\t"
// mtmp.transpose(pos);
"lqc2 vf7, 0x00(%1) \n\t"
"lqc2 $vf7, 0x00(%1) \n\t"
// m = mtmp.inverse();
"qmfc2.ni $11, vf0 \n\t"
"qmfc2.ni $8, vf4 \n\t"
"qmfc2.ni $9, vf5 \n\t"
"qmfc2.ni $10, vf6 \n\t"
"qmfc2.ni $11, $vf0 \n\t"
"qmfc2.ni $8, $vf4 \n\t"
"qmfc2.ni $9, $vf5 \n\t"
"qmfc2.ni $10, $vf6 \n\t"
"pextlw $12, $9, $8 \n\t"
"pextuw $13, $9, $8 \n\t"
@@ -318,18 +318,18 @@ void Matrix::setCamera(const float t_pos[4], const float t_vz[4], const float t_
"pcpyud $9, $12, $14 \n\t"
"pcpyld $10, $15, $13 \n\t"
"qmtc2.ni $8, vf16 \n\t"
"qmtc2.ni $9, vf17 \n\t"
"qmtc2.ni $10, vf18 \n\t"
"vmulax.xyz ACC, vf16, vf7 \n\t"
"vmadday.xyz ACC, vf17, vf7 \n\t"
"vmaddz.xyz vf5, vf18, vf7 \n\t"
"vsub.xyzw vf5, vf0, vf5 \n\t"
"qmtc2.ni $8, $vf16 \n\t"
"qmtc2.ni $9, $vf17 \n\t"
"qmtc2.ni $10, $vf18 \n\t"
"vmulax.xyz $ACC, $vf16, $vf7 \n\t"
"vmadday.xyz $ACC, $vf17, $vf7 \n\t"
"vmaddz.xyz $vf5, $vf18, $vf7 \n\t"
"vsub.xyzw $vf5, $vf0, $vf5 \n\t"
"sq $8, 0x00(%0) \n\t"
"sq $9, 0x10(%0) \n\t"
"sq $10, 0x20(%0) \n\t"
"sqc2 vf5, 0x30(%0) \n\t"
"sqc2 $vf5, 0x30(%0) \n\t"
:
: "r"(this->data), "r"(t_pos), "r"(t_vz), "r"(t_vy));
}
@@ -337,34 +337,34 @@ void Matrix::setCamera(const float t_pos[4], const float t_vz[4], const float t_
void Matrix::cross(float res[16], const float a[16], const float b[16]) const
{
asm volatile(
"lqc2 vf1, 0x00(%1) \n\t"
"lqc2 vf2, 0x10(%1) \n\t"
"lqc2 vf3, 0x20(%1) \n\t"
"lqc2 vf4, 0x30(%1) \n\t"
"lqc2 vf5, 0x00(%2) \n\t"
"lqc2 vf6, 0x10(%2) \n\t"
"lqc2 vf7, 0x20(%2) \n\t"
"lqc2 vf8, 0x30(%2) \n\t"
"vmulax.xyzw ACC, vf5, vf1 \n\t"
"vmadday.xyzw ACC, vf6, vf1 \n\t"
"vmaddaz.xyzw ACC, vf7, vf1 \n\t"
"vmaddw.xyzw vf1, vf8, vf1 \n\t"
"vmulax.xyzw ACC, vf5, vf2 \n\t"
"vmadday.xyzw ACC, vf6, vf2 \n\t"
"vmaddaz.xyzw ACC, vf7, vf2 \n\t"
"vmaddw.xyzw vf2, vf8, vf2 \n\t"
"vmulax.xyzw ACC, vf5, vf3 \n\t"
"vmadday.xyzw ACC, vf6, vf3 \n\t"
"vmaddaz.xyzw ACC, vf7, vf3 \n\t"
"vmaddw.xyzw vf3, vf8, vf3 \n\t"
"vmulax.xyzw ACC, vf5, vf4 \n\t"
"vmadday.xyzw ACC, vf6, vf4 \n\t"
"vmaddaz.xyzw ACC, vf7, vf4 \n\t"
"vmaddw.xyzw vf4, vf8, vf4 \n\t"
"sqc2 vf1, 0x00(%0) \n\t"
"sqc2 vf2, 0x10(%0) \n\t"
"sqc2 vf3, 0x20(%0) \n\t"
"sqc2 vf4, 0x30(%0) \n\t"
"lqc2 $vf1, 0x00(%1) \n\t"
"lqc2 $vf2, 0x10(%1) \n\t"
"lqc2 $vf3, 0x20(%1) \n\t"
"lqc2 $vf4, 0x30(%1) \n\t"
"lqc2 $vf5, 0x00(%2) \n\t"
"lqc2 $vf6, 0x10(%2) \n\t"
"lqc2 $vf7, 0x20(%2) \n\t"
"lqc2 $vf8, 0x30(%2) \n\t"
"vmulax.xyzw $ACC, $vf5, $vf1 \n\t"
"vmadday.xyzw $ACC, $vf6, $vf1 \n\t"
"vmaddaz.xyzw $ACC, $vf7, $vf1 \n\t"
"vmaddw.xyzw $vf1, $vf8, $vf1 \n\t"
"vmulax.xyzw $ACC, $vf5, $vf2 \n\t"
"vmadday.xyzw $ACC, $vf6, $vf2 \n\t"
"vmaddaz.xyzw $ACC, $vf7, $vf2 \n\t"
"vmaddw.xyzw $vf2, $vf8, $vf2 \n\t"
"vmulax.xyzw $ACC, $vf5, $vf3 \n\t"
"vmadday.xyzw $ACC, $vf6, $vf3 \n\t"
"vmaddaz.xyzw $ACC, $vf7, $vf3 \n\t"
"vmaddw.xyzw $vf3, $vf8, $vf3 \n\t"
"vmulax.xyzw $ACC, $vf5, $vf4 \n\t"
"vmadday.xyzw $ACC, $vf6, $vf4 \n\t"
"vmaddaz.xyzw $ACC, $vf7, $vf4 \n\t"
"vmaddw.xyzw $vf4, $vf8, $vf4 \n\t"
"sqc2 $vf1, 0x00(%0) \n\t"
"sqc2 $vf2, 0x10(%0) \n\t"
"sqc2 $vf3, 0x20(%0) \n\t"
"sqc2 $vf4, 0x30(%0) \n\t"
:
: "r"(res), "r"(b), "r"(a)
: "memory");
src/engine/models/math/vector3.cpp
+80 -80
View File
@@ -30,10 +30,10 @@ Vector3 Vector3::operator+(const Vector3 &v) const
{
Vector3 result;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t"
"lqc2 vf5, 0x0(%2) \n\t"
"vadd.xyz vf6, vf4, vf5 \n\t"
"sqc2 vf6, 0x0(%0) \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"lqc2 $vf5, 0x0(%2) \n\t"
"vadd.xyz $vf6, $vf4, $vf5 \n\t"
"sqc2 $vf6, 0x0(%0) \n\t"
:
: "r"(result.xyz), "r"(this->xyz), "r"(v.xyz));
return result;
@@ -43,10 +43,10 @@ Vector3 Vector3::operator-(const Vector3 &v) const
{
Vector3 result;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t"
"lqc2 vf5, 0x0(%2) \n\t"
"vsub.xyz vf6, vf4, vf5 \n\t"
"sqc2 vf6, 0x0(%0) \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"lqc2 $vf5, 0x0(%2) \n\t"
"vsub.xyz $vf6, $vf4, $vf5 \n\t"
"sqc2 $vf6, 0x0(%0) \n\t"
:
: "r"(result.xyz), "r"(this->xyz), "r"(v.xyz));
return result;
@@ -55,13 +55,13 @@ Vector3 Vector3::operator-(const Vector3 &v) const
Vector3 Vector3::operator*(const Vector3 &v) const
{
Vector3 res;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t" // vf4 = this
"lqc2 vf5, 0x0(%2) \n\t" // vf5 = v
"vopmula.xyz ACC, vf4, vf5 \n\t"
"vopmsub.xyz vf8, vf5, vf4 \n\t"
"vsub.w vf8, vf00, vf00 \n\t"
"sqc2 vf8, 0x0(%0) \n\t" // vf8 = res
asm volatile( // VU0 Macro program
"lqc2 $vf4, 0x0(%1) \n\t" // $vf4 = this
"lqc2 $vf5, 0x0(%2) \n\t" // $vf5 = v
"vopmula.xyz $ACC, $vf4, $vf5 \n\t"
"vopmsub.xyz $vf8, $vf5, $vf4 \n\t"
"vsub.w $vf8, $vf0, $vf0 \n\t"
"sqc2 $vf8, 0x0(%0) \n\t" // $vf8 = res
:
: "r"(res.xyz), "r"(this->xyz), "r"(v.xyz));
// result.x = y * v.z - z * v.y;
@@ -74,11 +74,11 @@ Vector3 Vector3::operator*(const float &t) const
{
Vector3 result;
asm volatile(
"lqc2 vf4, 0x0(%1) \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"mfc1 $8, %2 \n\t"
"qmtc2 $8, vf5 \n\t"
"vmulx.xyz vf6, vf4, vf5 \n\t"
"sqc2 vf6, 0x0(%0) \n\t"
"qmtc2 $8, $vf5 \n\t"
"vmulx.xyz $vf6, $vf4, $vf5 \n\t"
"sqc2 $vf6, 0x0(%0) \n\t"
:
: "r"(result.xyz), "r"(this->xyz), "f"(t));
return result;
@@ -96,10 +96,10 @@ Vector3 Vector3::operator/(const float &t) const
void Vector3::operator+=(const Vector3 &t)
{
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%0) \n\t"
"lqc2 vf5, 0x0(%1) \n\t"
"vadd.xyz vf4, vf4, vf5 \n\t"
"sqc2 vf4, 0x0(%0) \n\t"
"lqc2 $vf4, 0x0(%0) \n\t"
"lqc2 $vf5, 0x0(%1) \n\t"
"vadd.xyz $vf4, $vf4, $vf5 \n\t"
"sqc2 $vf4, 0x0(%0) \n\t"
:
: "r"(this->xyz), "r"(t.xyz));
}
@@ -107,11 +107,11 @@ void Vector3::operator+=(const Vector3 &t)
void Vector3::operator*=(const float &t)
{
asm volatile(
"lqc2 vf4, 0x0(%0) \n\t"
"lqc2 $vf4, 0x0(%0) \n\t"
"mfc1 $8, %1 \n\t"
"qmtc2 $8, vf5 \n\t"
"vmulx.xyz vf4, vf4, vf5 \n\t"
"sqc2 vf4, 0x0(%0) \n\t"
"qmtc2 $8, $vf5 \n\t"
"vmulx.xyz $vf4, $vf4, $vf5 \n\t"
"sqc2 $vf4, 0x0(%0) \n\t"
:
: "r"(this->xyz), "f"(t));
}
@@ -140,12 +140,12 @@ float Vector3::innerProduct(const Vector3 &v) const
{
float result;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t"
"lqc2 vf5, 0x0(%2) \n\t"
"vmul.xyz vf6, vf4, vf5 \n\t"
"vaddy.x vf6, vf6, vf6 \n\t"
"vaddz.x vf6, vf6, vf6 \n\t"
"qmfc2 $2, vf6 \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"lqc2 $vf5, 0x0(%2) \n\t"
"vmul.xyz $vf6, $vf4, $vf5 \n\t"
"vaddy.x $vf6, $vf6, $vf6 \n\t"
"vaddz.x $vf6, $vf6, $vf6 \n\t"
"qmfc2 $2, $vf6 \n\t"
"mtc1 $2, %0 \n\t"
: "=f"(result)
: "r"(this->xyz), "r"(v.xyz));
@@ -157,14 +157,14 @@ float Vector3::length() const
{
float result;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t"
"vmul.xyz vf5, vf4, vf4 \n\t"
"vaddy.x vf5, vf5, vf5 \n\t"
"vaddz.x vf5, vf5, vf5 \n\t"
"vsqrt Q , vf5x \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"vmul.xyz $vf5, $vf4, $vf4 \n\t"
"vaddy.x $vf5, $vf5, $vf5 \n\t"
"vaddz.x $vf5, $vf5, $vf5 \n\t"
"vsqrt $Q , $vf5x \n\t"
"vwaitq \n\t"
"vaddq.x vf8, vf0, Q \n\t"
"qmfc2 $2, vf8 \n\t"
"vaddq.x $vf8, $vf0, $Q \n\t"
"qmfc2 $2, $vf8 \n\t"
"mtc1 $2, %0 \n\t"
: "=f"(result)
: "r"(this->xyz));
@@ -175,17 +175,17 @@ float Vector3::length() const
void Vector3::normalize()
{
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%0) \n\t"
"vmul.xyz vf5, vf4, vf4 \n\t"
"vaddy.x vf5, vf5, vf5 \n\t"
"vaddz.x vf5, vf5, vf5 \n\t"
"vrsqrt Q, vf0w, vf5x \n\t"
"lqc2 $vf4, 0x0(%0) \n\t"
"vmul.xyz $vf5, $vf4, $vf4 \n\t"
"vaddy.x $vf5, $vf5, $vf5 \n\t"
"vaddz.x $vf5, $vf5, $vf5 \n\t"
"vrsqrt $Q, $vf0w, $vf5x \n\t"
"vwaitq \n\t"
"vsub.xyz vf6, vf0, vf0 \n\t"
"vaddw.xyz vf6, vf6, vf4 \n\t"
"vsub.xyz $vf6, $vf0, $vf0 \n\t"
"vaddw.xyz $vf6, $vf6, $vf4 \n\t"
"vwaitq \n\t"
"vmulq.xyz vf6, vf4, Q \n\t"
"sqc2 vf6, 0x0(%0) \n\t"
"vmulq.xyz $vf6, $vf4, $Q \n\t"
"sqc2 $vf6, 0x0(%0) \n\t"
:
: "r"(this->xyz));
}
@@ -194,16 +194,16 @@ float Vector3::distanceTo(const Vector3 &v) const
{
register float result;
asm volatile( // VU0 Macro program
"lqc2 vf4, 0x0(%1) \n\t"
"lqc2 vf5, 0x0(%2) \n\t"
"vsub.xyz vf6, vf4, vf5 \n\t"
"vmul.xyz vf7, vf6, vf6 \n\t"
"vaddy.x vf7, vf7, vf7 \n\t"
"vaddz.x vf7, vf7, vf7 \n\t"
"vsqrt Q , vf7x \n\t"
"lqc2 $vf4, 0x0(%1) \n\t"
"lqc2 $vf5, 0x0(%2) \n\t"
"vsub.xyz $vf6, $vf4, $vf5 \n\t"
"vmul.xyz $vf7, $vf6, $vf6 \n\t"
"vaddy.x $vf7, $vf7, $vf7 \n\t"
"vaddz.x $vf7, $vf7, $vf7 \n\t"
"vsqrt $Q , $vf7x \n\t"
"vwaitq \n\t"
"vaddq.x vf8, vf0, Q \n\t"
"qmfc2 $2, vf8 \n\t"
"vaddq.x $vf8, $vf0, $Q \n\t"
"qmfc2 $2, $vf8 \n\t"
"mtc1 $2, %0 \n\t"
: "=f"(result)
: "r"(this->xyz), "r"(v.xyz));
@@ -217,20 +217,20 @@ u8 Vector3::shouldBeBackfaceCulled(const Vector3 *t_cameraPos, const Vector3 *t_
{
register float dot;
asm volatile(
"lqc2 vf4, 0x0(%1) \n\t" // vf4 = cameraPos
"lqc2 vf5, 0x0(%2) \n\t" // vf5 = v0
"lqc2 vf6, 0x0(%3) \n\t" // vf6 = v1
"lqc2 vf7, 0x0(%4) \n\t" // vf7 = v2
"vsub.xyz vf8, vf7, vf5 \n\t" // vf8 = vf7(v2) - vf5(v0)
"vsub.xyz vf9, vf6, vf5 \n\t" // vf9 = vf6(v1) - vf5(v0)
"vopmula.xyz ACC, vf8, vf9 \n\t" // vf6 = cross(vf8, vf9)
"vopmsub.xyz vf6, vf9, vf8 \n\t"
"vsub.w vf6, vf6, vf6 \n\t"
"vsub.xyz vf7, vf5, vf4 \n\t" // vf7 = vf5(v0) - vf4(cameraPos)
"vmul.xyz vf5, vf7, vf6 \n\t" // vf5 = dot(vf7, vf6)
"vaddy.x vf5, vf5, vf5 \n\t"
"vaddz.x vf5, vf5, vf5 \n\t"
"qmfc2 $2, vf5 \n\t" // store result on `dot` variable
"lqc2 $vf4, 0x0(%1) \n\t" // $vf4 = cameraPos
"lqc2 $vf5, 0x0(%2) \n\t" // $vf5 = v0
"lqc2 $vf6, 0x0(%3) \n\t" // $vf6 = v1
"lqc2 $vf7, 0x0(%4) \n\t" // $vf7 = v2
"vsub.xyz $vf8, $vf7, $vf5 \n\t" // $vf8 = $vf7(v2) - $vf5(v0)
"vsub.xyz $vf9, $vf6, $vf5 \n\t" // $vf9 = $vf6(v1) - $vf5(v0)
"vopmula.xyz $ACC, $vf8, $vf9 \n\t" // $vf6 = cross($vf8, $vf9)
"vopmsub.xyz $vf6, $vf9, $vf8 \n\t"
"vsub.w $vf6, $vf6, $vf6 \n\t"
"vsub.xyz $vf7, $vf5, $vf4 \n\t" // $vf7 = $vf5(v0) - $vf4(cameraPos)
"vmul.xyz $vf5, $vf7, $vf6 \n\t" // $vf5 = dot($vf7, $vf6)
"vaddy.x $vf5, $vf5, $vf5 \n\t"
"vaddz.x $vf5, $vf5, $vf5 \n\t"
"qmfc2 $2, $vf5 \n\t" // store result on `dot` variable
"mtc1 $2, %0 \n\t"
: "=f"(dot)
: "r"(t_cameraPos->xyz), "r"(t_v0->xyz), "r"(t_v1->xyz), "r"(t_v2->xyz));
@@ -240,14 +240,14 @@ u8 Vector3::shouldBeBackfaceCulled(const Vector3 *t_cameraPos, const Vector3 *t_
void Vector3::setByLerp(const Vector3 &t_v1, const Vector3 &t_v2, const float &t_interp, const float &t_scale)
{
asm volatile(
"lqc2 vf4, 0x0(%1) \n\t" // vf4 = v1
"lqc2 vf5, 0x0(%2) \n\t" // vf5 = v2
"mfc1 $8, %3 \n\t" // vf6 = t
"qmtc2 $8, vf6 \n\t" // lerp:
"vsub.xyz vf7, vf5, vf4 \n\t" // vf7 = v2 - v1
"vmulx.xyz vf8, vf7, vf6 \n\t" // vf8 = vf7 * t
"vadd.xyz vf9, vf8, vf4 \n\t" // vf9 = vf8 + vf4
"sqc2 vf9, 0x0(%0) \n\t" // v0 = vf9
"lqc2 $vf4, 0x0(%1) \n\t" // $vf4 = v1
"lqc2 $vf5, 0x0(%2) \n\t" // $vf5 = v2
"mfc1 $8, %3 \n\t" // $vf6 = t
"qmtc2 $8, $vf6 \n\t" // lerp:
"vsub.xyz $vf7, $vf5, $vf4 \n\t" // $vf7 = v2 - v1
"vmulx.xyz $vf8, $vf7, $vf6 \n\t" // $vf8 = $vf7 * t
"vadd.xyz $vf9, $vf8, $vf4 \n\t" // $vf9 = $vf8 + $vf4
"sqc2 $vf9, 0x0(%0) \n\t" // v0 = $vf9
:
: "r"(&this->xyz), "r"(&t_v1.xyz), "r"(&t_v2.xyz), "f"(t_interp));
operator*=(t_scale);
src/engine/models/mesh.cpp
+58 -58
View File
@@ -195,7 +195,7 @@ u32 Mesh::getDrawData(u32 t_materialIndex, VECTOR *o_vertices, VECTOR *o_normals
asm volatile(
// VU0 macro program:
// Load vector with 1.0F values to VF21
"lqc2 vf21, 0x0(%0) \n\t" // load "one vec"
"lqc2 $vf21, 0x0(%0) \n\t" // load "one vec"
:
: "r"(ONE_VEC));
@@ -221,34 +221,34 @@ u32 Mesh::getDrawData(u32 t_materialIndex, VECTOR *o_vertices, VECTOR *o_normals
// Calculate lerp() and store data into calc3Vectors
// Vertex 0
"lqc2 vf4, 0x0(%3) \n\t" // vf4 = v1
"lqc2 vf5, 0x0(%6) \n\t" // vf5 = v2
"mfc1 $10, %9 \n\t" // vf6 = t
"qmtc2 $10, vf6 \n\t" // lerp:
"vsub.xyz vf7, vf5, vf4 \n\t" // vf7 = v2 - v1
"vmulx.xyz vf8, vf7, vf6 \n\t" // vf8 = vf7 * t
"vadd.xyz vf9, vf8, vf4 \n\t" // vf9 = vf8 + vf4
"sqc2 vf9, 0x0(%0) \n\t" // v0 = vf9
"lqc2 $vf4, 0x0(%3) \n\t" // $vf4 = v1
"lqc2 $vf5, 0x0(%6) \n\t" // $vf5 = v2
"mfc1 $10, %9 \n\t" // $vf6 = t
"qmtc2 $10, $vf6 \n\t" // lerp:
"vsub.xyz $vf7, $vf5, $vf4 \n\t" // $vf7 = v2 - v1
"vmulx.xyz $vf8, $vf7, $vf6 \n\t" // $vf8 = $vf7 * t
"vadd.xyz $vf9, $vf8, $vf4 \n\t" // $vf9 = $vf8 + $vf4
"sqc2 $vf9, 0x0(%0) \n\t" // v0 = $vf9
// Vertex 1
"lqc2 vf4, 0x0(%4) \n\t" // vf4 = v1
"lqc2 vf5, 0x0(%7) \n\t" // vf5 = v2
"mfc1 $10, %9 \n\t" // vf6 = t
"qmtc2 $10, vf6 \n\t" // lerp:
"vsub.xyz vf7, vf5, vf4 \n\t" // vf7 = v2 - v1
"vmulx.xyz vf8, vf7, vf6 \n\t" // vf8 = vf7 * t
"vadd.xyz vf9, vf8, vf4 \n\t" // vf9 = vf8 + vf4
"sqc2 vf9, 0x0(%1) \n\t" // v0 = vf9
"lqc2 $vf4, 0x0(%4) \n\t" // $vf4 = v1
"lqc2 $vf5, 0x0(%7) \n\t" // $vf5 = v2
"mfc1 $10, %9 \n\t" // $vf6 = t
"qmtc2 $10, $vf6 \n\t" // lerp:
"vsub.xyz $vf7, $vf5, $vf4 \n\t" // $vf7 = v2 - v1
"vmulx.xyz $vf8, $vf7, $vf6 \n\t" // $vf8 = $vf7 * t
"vadd.xyz $vf9, $vf8, $vf4 \n\t" // $vf9 = $vf8 + $vf4
"sqc2 $vf9, 0x0(%1) \n\t" // v0 = $vf9
// Vertex 2
"lqc2 vf4, 0x0(%5) \n\t" // vf4 = v1
"lqc2 vf5, 0x0(%8) \n\t" // vf5 = v2
"mfc1 $10, %9 \n\t" // vf6 = t
"qmtc2 $10, vf6 \n\t" // lerp:
"vsub.xyz vf7, vf5, vf4 \n\t" // vf7 = v2 - v1
"vmulx.xyz vf8, vf7, vf6 \n\t" // vf8 = vf7 * t
"vadd.xyz vf9, vf8, vf4 \n\t" // vf9 = vf8 + vf4
"sqc2 vf9, 0x0(%2) \n\t" // v0 = vf9
"lqc2 $vf4, 0x0(%5) \n\t" // $vf4 = v1
"lqc2 $vf5, 0x0(%8) \n\t" // $vf5 = v2
"mfc1 $10, %9 \n\t" // $vf6 = t
"qmtc2 $10, $vf6 \n\t" // lerp:
"vsub.xyz $vf7, $vf5, $vf4 \n\t" // $vf7 = v2 - v1
"vmulx.xyz $vf8, $vf7, $vf6 \n\t" // $vf8 = $vf7 * t
"vadd.xyz $vf9, $vf8, $vf4 \n\t" // $vf9 = $vf8 + $vf4
"sqc2 $vf9, 0x0(%2) \n\t" // v0 = $vf9
:
: "r"(calc3Vectors[0].xyz),
@@ -267,12 +267,12 @@ u32 Mesh::getDrawData(u32 t_materialIndex, VECTOR *o_vertices, VECTOR *o_normals
asm volatile(
// VU0 macro program
// Copy 0,1,2 vertices
"lqc2 vf1, 0x0(%3) \n\t" // load vert
"lqc2 vf2, 0x0(%4) \n\t" // load normal
"lqc2 vf3, 0x0(%5) \n\t" // load st
"sqc2 vf1, 0x0(%0) \n\t" // store vert
"sqc2 vf2, 0x0(%1) \n\t" // store normal
"sqc2 vf3, 0x0(%2) \n\t" // store st
"lqc2 $vf1, 0x0(%3) \n\t" // load vert
"lqc2 $vf2, 0x0(%4) \n\t" // load normal
"lqc2 $vf3, 0x0(%5) \n\t" // load st
"sqc2 $vf1, 0x0(%0) \n\t" // store vert
"sqc2 $vf2, 0x0(%1) \n\t" // store normal
"sqc2 $vf3, 0x0(%2) \n\t" // store st
:
: "r"(calc3Vectors[0].xyz),
"r"(calc3Vectors[1].xyz),
@@ -291,37 +291,37 @@ u32 Mesh::getDrawData(u32 t_materialIndex, VECTOR *o_vertices, VECTOR *o_normals
// Copy data and set vert/normal "w" and st "z"+"w" to 1.0F
// Vertex 0
"lqc2 vf1, 0x0(%3) \n\t" // load vert
"lqc2 vf2, 0x0(%4) \n\t" // load normal
"lqc2 vf3, 0x0(%5) \n\t" // load st
"vadd.w vf1, vf20, vf21 \n\t" // set vert.w to 1.0F
"vadd.w vf2, vf20, vf21 \n\t" // set normal.W to 1.0F
"vadd.zw vf3, vf20, vf21 \n\t" // set st.zw to 1.0F
"sqc2 vf1, 0x0(%0) \n\t" // store vert
"sqc2 vf2, 0x0(%1) \n\t" // store normal
"sqc2 vf3, 0x0(%2) \n\t" // store st
"lqc2 $vf1, 0x0(%3) \n\t" // load vert
"lqc2 $vf2, 0x0(%4) \n\t" // load normal
"lqc2 $vf3, 0x0(%5) \n\t" // load st
"vadd.w $vf1, $vf20, $vf21 \n\t" // set vert.w to 1.0F
"vadd.w $vf2, $vf20, $vf21 \n\t" // set normal.W to 1.0F
"vadd.zw $vf3, $vf20, $vf21 \n\t" // set st.zw to 1.0F
"sqc2 $vf1, 0x0(%0) \n\t" // store vert
"sqc2 $vf2, 0x0(%1) \n\t" // store normal
"sqc2 $vf3, 0x0(%2) \n\t" // store st
// Vertex 1
"lqc2 vf1, 0x0(%9) \n\t" // load vert
"lqc2 vf2, 0x0(%10) \n\t" // load normal
"lqc2 vf3, 0x0(%11) \n\t" // load st
"vadd.w vf1, vf20, vf21 \n\t" // set vert.w to 1.0F
"vadd.w vf2, vf20, vf21 \n\t" // set normal.W to 1.0F
"vadd.zw vf3, vf20, vf21 \n\t" // set st.zw to 1.0F
"sqc2 vf1, 0x0(%6) \n\t" // store vert
"sqc2 vf2, 0x0(%7) \n\t" // store normal
"sqc2 vf3, 0x0(%8) \n\t" // store st
"lqc2 $vf1, 0x0(%9) \n\t" // load vert
"lqc2 $vf2, 0x0(%10) \n\t" // load normal
"lqc2 $vf3, 0x0(%11) \n\t" // load st
"vadd.w $vf1, $vf20, $vf21 \n\t" // set vert.w to 1.0F
"vadd.w $vf2, $vf20, $vf21 \n\t" // set normal.W to 1.0F
"vadd.zw $vf3, $vf20, $vf21 \n\t" // set st.zw to 1.0F
"sqc2 $vf1, 0x0(%6) \n\t" // store vert
"sqc2 $vf2, 0x0(%7) \n\t" // store normal
"sqc2 $vf3, 0x0(%8) \n\t" // store st
// Vertex 2
"lqc2 vf1, 0x0(%15) \n\t" // load vert
"lqc2 vf2, 0x0(%16) \n\t" // load normal
"lqc2 vf3, 0x0(%17) \n\t" // load st
"vadd.w vf1, vf20, vf21 \n\t" // set vert.w to 1.0F
"vadd.w vf2, vf20, vf21 \n\t" // set normal.W to 1.0F
"vadd.zw vf3, vf20, vf21 \n\t" // set st.zw to 1.0F
"sqc2 vf1, 0x0(%12) \n\t" // store vert
"sqc2 vf2, 0x0(%13) \n\t" // store normal
"sqc2 vf3, 0x0(%14) \n\t" // store st
"lqc2 $vf1, 0x0(%15) \n\t" // load vert
"lqc2 $vf2, 0x0(%16) \n\t" // load normal
"lqc2 $vf3, 0x0(%17) \n\t" // load st
"vadd.w $vf1, $vf20, $vf21 \n\t" // set vert.w to 1.0F
"vadd.w $vf2, $vf20, $vf21 \n\t" // set normal.W to 1.0F
"vadd.zw $vf3, $vf20, $vf21 \n\t" // set st.zw to 1.0F
"sqc2 $vf1, 0x0(%12) \n\t" // store vert
"sqc2 $vf2, 0x0(%13) \n\t" // store normal
"sqc2 $vf3, 0x0(%14) \n\t" // store st
:
: "r"(o_vertices[addedFaces]),