diff --git a/spec b/spec
index da9ea49997..8ec349fd98 100644
--- a/spec
+++ b/spec
@@ -155,7 +155,6 @@ beginseg
     include "build/src/libultra/io/epirawread.o"
     include "build/src/libultra/io/viswapbuf.o"
     include "build/src/libultra/gu/position.o"
-    include "build/data/boot/position.data.o"
     include "build/src/libultra/io/epirawdma.o"
     include "build/src/libultra/io/sptaskyielded.o"
     include "build/asm/boot/bcmp.text.o"
diff --git a/src/libultra/gu/position.c b/src/libultra/gu/position.c
index 2892b1f11e..589d4f4a31 100644
--- a/src/libultra/gu/position.c
+++ b/src/libultra/gu/position.c
@@ -1,5 +1,54 @@
 #include "global.h"
 
-#pragma GLOBAL_ASM("asm/non_matchings/boot/position/guPositionF.s")
+/**
+ * guPositionF
+ * Creates a rotation/parallel translation modelling matrix (floating point)
+ */
+void guPositionF(f32 mf[4][4], f32 rot, f32 pitch, f32 yaw, f32 scale, f32 x, f32 y, f32 z) {
+    static f32 D_80134D00 = M_PI / 180.0;
+    f32 sinr, sinp, sinh;
+    f32 cosr, cosp, cosh;
 
-#pragma GLOBAL_ASM("asm/non_matchings/boot/position/guPosition.s")
+    rot *= D_80134D00;
+    pitch *= D_80134D00;
+    yaw *= D_80134D00;
+
+    sinr = __sinf(rot);
+    cosr = __cosf(rot);
+    sinp = __sinf(pitch);
+    cosp = __cosf(pitch);
+    sinh = __sinf(yaw);
+    cosh = __cosf(yaw);
+
+    mf[0][0] = (cosp * cosh) * scale;
+    mf[0][1] = (cosp * sinh) * scale;
+    mf[0][2] = (-sinp) * scale;
+    mf[0][3] = 0.0f;
+
+    mf[1][0] = ((sinr * sinp * cosh) - (cosr * sinh)) * scale;
+    mf[1][1] = ((sinr * sinp * sinh) + (cosr * cosh)) * scale;
+    mf[1][2] = (sinr * cosp) * scale;
+    mf[1][3] = 0.0f;
+
+    mf[2][0] = ((cosr * sinp * cosh) + (sinr * sinh)) * scale;
+    mf[2][1] = ((cosr * sinp * sinh) - (sinr * cosh)) * scale;
+    mf[2][2] = (cosr * cosp) * scale;
+    mf[2][3] = 0.0f;
+
+    mf[3][0] = x;
+    mf[3][1] = y;
+    mf[3][2] = z;
+    mf[3][3] = 1.0f;
+}
+
+/**
+ * guPosition
+ * Creates a rotational/parallel translation modeling matrix (fixed point)
+ */
+void guPosition(Mtx* m, f32 rot, f32 pitch, f32 yaw, f32 scale, f32 x, f32 y, f32 z) {
+    f32 mf[4][4];
+
+    guPositionF(mf, rot, pitch, yaw, scale, x, y, z);
+
+    guMtxF2L(mf, m);
+}