diff --git a/examples/PoolBox/application/src/Overlay/GTETest/gte_test.cpp b/examples/PoolBox/application/src/Overlay/GTETest/gte_test.cpp
index 368454fd..8d860602 100644
--- a/examples/PoolBox/application/src/Overlay/GTETest/gte_test.cpp
+++ b/examples/PoolBox/application/src/Overlay/GTETest/gte_test.cpp
@@ -30,7 +30,7 @@ namespace GTETest {
             return true;
         }
 
-        auto matrix = GTE::MATRIX::rotation(0, rotation, rotation);
+        auto matrix = GTE::MATRIX::rotation(0, 0, rotation);
         matrix.trans[0] = (Assets::Main::DoenerFishInfo.size.width/2);
         matrix.trans[1] = (Assets::Main::DoenerFishInfo.size.height/2);
 
diff --git a/examples/PoolBox/application/src/Overlay/GTETest/include/GTE_Sprite.hpp b/examples/PoolBox/application/src/Overlay/GTETest/include/GTE_Sprite.hpp
index 44121684..256ec8fa 100644
--- a/examples/PoolBox/application/src/Overlay/GTETest/include/GTE_Sprite.hpp
+++ b/examples/PoolBox/application/src/Overlay/GTETest/include/GTE_Sprite.hpp
@@ -27,7 +27,8 @@ namespace GTETest {
                 move_back.trans[1] = -matrix.trans[1];
                 move_back.trans[2] = -matrix.trans[2];
 
-                return GTE::apply_matrix(matrix, GTE::apply_matrix(move_back, vector, output), output).to<GPU::Vertex>();
+                GTE::comp_matrix(matrix, move_back, move_back);
+                return GTE::apply_matrix(move_back, vector, output).to<GPU::Vertex>();
             };
 
             const auto& area = this->area;
diff --git a/include/PSX/GTE/gte.hpp b/include/PSX/GTE/gte.hpp
index 2bedda8d..e911657a 100644
--- a/include/PSX/GTE/gte.hpp
+++ b/include/PSX/GTE/gte.hpp
@@ -115,6 +115,24 @@ namespace JabyEngine {
         */
         MATRIX& multiply_matrix(const MATRIX& m0, const MATRIX& m1, MATRIX& result);
 
+        /*
+            CompMatrix
+
+            m0: first input
+            m1: second input
+            result: result of computing m0 and m1
+            return: returns result
+        */
+        static MATRIX& comp_matrix(const MATRIX& m0, const MATRIX& m1, MATRIX& result) {
+            multiply_matrix(m0, m1, result);
+            set_trans_matrix(m0);
+            GTE::ldlv0(reinterpret_cast<const VECTOR&>(m1.trans));
+            GTE::rt();
+            GTE::stlvnl(reinterpret_cast<VECTOR&>(result.trans));
+
+			return result;
+        } 
+
         /*
             matrix: first input
 
diff --git a/include/PSX/GTE/gte_instruction.hpp b/include/PSX/GTE/gte_instruction.hpp
index 0cc51473..44cb42d6 100644
--- a/include/PSX/GTE/gte_instruction.hpp
+++ b/include/PSX/GTE/gte_instruction.hpp
@@ -21,6 +21,17 @@ namespace JabyEngine {
             __asm__ volatile("lwc2 $5, 4(%0)" :: "r"(&vector));
         }
 
+        // Load LS 16 bits of VECTOR to 16 bit universal vector.
+        static __always_inline void ldlv0(const VECTOR& vector) {
+            __asm__ volatile("lhu  $13, 4(%0)"    :: "r"(&vector) : "$12", "$13");
+            __asm__ volatile("lhu  $12, 0(%0)"    :: "r"(&vector) : "$12", "$13");
+            __asm__ volatile("sll  $13, $13, 16"  :: "r"(&vector) : "$12", "$13");
+            __asm__ volatile("or   $12, $12, $13" :: "r"(&vector) : "$12", "$13");
+            __asm__ volatile("mtc2 $12, $0"       :: "r"(&vector) : "$12", "$13");
+            __asm__ volatile("lwc2 $1, 8(%0)"     :: "r"(&vector) : "$12", "$13");
+        
+        }
+
         // Load column vector of MATRIX to universal register
         static __always_inline void ldclmv(const MATRIX& matrix, size_t col) {
             __asm__ volatile("lhu  $12, 0(%0)"  :: "r"(reinterpret_cast<uintptr_t>(&matrix) + (col << 1)) : "$12", "$13", "$14");