diff --git a/compiler/examples/cube3d-c64.p8 b/compiler/examples/cube3d-c64.p8
index 31783e11d..6c4f442ff 100644
--- a/compiler/examples/cube3d-c64.p8
+++ b/compiler/examples/cube3d-c64.p8
@@ -1,4 +1,5 @@
 %import c64utils
+%import c64flt
 
 ~ main {
 
@@ -6,104 +7,89 @@
     const uword height = 25
 
     ; vertices
-    byte[8] xcoor = [ -100, -100, -100, -100,  100,  100,  100, 100 ]
-    byte[8] ycoor = [ -100, -100,  100,  100, -100, -100,  100, 100 ]
-    byte[8] zcoor = [ -100,  100, -100,  100, -100,  100, -100, 100 ]
+    byte[8] xcoor = [ -50, -50, -50, -50,  50,  50,  50, 50 ]
+    byte[8] ycoor = [ -50, -50,  50,  50, -50, -50,  50, 50 ]
+    byte[8] zcoor = [ -50,  50, -50,  50, -50,  50, -50, 50 ]
 
     ; storage for rotated coordinates
-    word[len(xcoor)] rotatedx=0
-    word[len(ycoor)] rotatedy=0
-    word[len(zcoor)] rotatedz=-32767
+    float[len(xcoor)] rotatedx=0.0
+    float[len(ycoor)] rotatedy=0.0
+    float[len(zcoor)] rotatedz=-1.0
 
     sub start()  {
         uword anglex
         uword angley
         uword anglez
         while(true) {
-            rotate_vertices(anglex, angley, anglez)
-            ; c64.CLEARSCR()
-            ; draw_edges()
-            anglex+=256
-            angley+=83
-            anglez+=201
+            rotate_vertices(msb(anglex), msb(angley), msb(anglez))
+            c64.CLEARSCR()
+            draw_edges()
+            anglex+=1000
+            angley+=333
+            anglez+=807
         }
     }
 
-    sub rotate_vertices(uword ax, uword ay, uword az) {
+    sub rotate_vertices(ubyte ax, ubyte ay, ubyte az) {
         ; rotate around origin (0,0,0)
 
         ; set up the 3d rotation matrix values
-        word cosa = cos8(msb(ax)) as word
-        word sina = sin8(msb(ax)) as word
-        word cosb = cos8(msb(ay)) as word
-        word sinb = sin8(msb(ay)) as word
-        word cosc = cos8(msb(az)) as word
-        word sinc = sin8(msb(az)) as word
+        float cosa = cos8(ax) as float / 128.0
+        float sina = sin8(ax) as float / 128.0
+        float cosb = cos8(ay) as float / 128.0
+        float sinb = sin8(ay) as float / 128.0
+        float cosc = cos8(az) as float / 128.0
+        float sinc = sin8(az) as float / 128.0
 
-        word cosa_sinb = msb(cosa*sinb)
-        word sina_sinb = msb(sina*sinb)
-
-        c64.CHROUT('>')
-        c64scr.print_w(cosa_sinb)
-        c64.CHROUT(',')
-        c64scr.print_w(sina_sinb)
-        c64.CHROUT('\n')
-
-        word Axx = msb(cosa*cosb)
-        word Axy = msb(cosa_sinb*sinc - sina*cosc)
-        word Axz = msb(cosa_sinb*cosc + sina*sinc)
-        word Ayx = msb(sina*cosb)
-        word Ayy = msb(sina_sinb*sinc + cosa*cosc)
-        word Ayz = msb(sina_sinb*cosc - cosa*sinc)
-        word Azx = -sinb
-        word Azy = msb(cosb*sinc)
-        word Azz = msb(cosb*cosc)
-
-        c64.CHROUT('>')
-        c64scr.print_w(Axx)
-        c64.CHROUT(',')
-        c64scr.print_w(Axy)
-        c64.CHROUT(',')
-        c64scr.print_w(Axz)
-        c64.CHROUT('\n')
+        float cosa_sinb = cosa*sinb
+        float sina_sinb = sina*sinb
+        float Axx = cosa*cosb
+        float Axy = cosa_sinb*sinc - sina*cosc
+        float Axz = cosa_sinb*cosc + sina*sinc
+        float Ayx = sina*cosb
+        float Ayy = sina_sinb*sinc + cosa*cosc
+        float Ayz = sina_sinb*cosc - cosa*sinc
+        float Azx = -sinb
+        float Azy = cosb*sinc
+        float Azz = cosb*cosc
 
         for ubyte i in 0 to len(xcoor)-1 {
-            word xc = xcoor[i]
-            word yc = ycoor[i]
-            word zc = zcoor[i]
-            rotatedx[i] = Axx*xc ;+ Axy*yc + Axz*zc      ; @todo wrong code generated? crash!
-            rotatedy[i] = Ayx*xc ;+ Ayy*yc + Ayz*zc      ; @todo wrong code generated? crash!
-            rotatedz[i] = Azx*xc ;+ Azy*yc + Azz*zc      ; @todo wrong code generated? crash!
+            float xc = xcoor[i] as float
+            float yc = ycoor[i] as float
+            float zc = zcoor[i] as float
+            rotatedx[i] = Axx*xc + Axy*yc + Axz*zc
+            rotatedy[i] = Ayx*xc + Ayy*yc + Ayz*zc
+            rotatedz[i] = Azx*xc + Azy*yc + Azz*zc
         }
     }
 
     sub draw_edges() {
 
-        sub toscreenx(word x, word z) -> ubyte {
-            return msb(x) and 31
+        sub toscreenx(float x, float z) -> byte {
+            return x/(250.0+z) * (height as float) as byte + width // 2
         }
 
-        sub toscreeny(word y, word z) -> ubyte {
-            return msb(y) and 15
+        sub toscreeny(float y, float z) -> byte {
+            return y/(250.0+z) * (height as float) as byte + height // 2
         }
 
         ; plot the points of the 3d cube
         ; first the points on the back, then the points on the front (painter algorithm)
 
         for ubyte i in 0 to len(xcoor)-1 {
-            word rz = rotatedz[i]
-            if rz >= 100 {
-                ubyte sx = toscreenx(rotatedx[i], rz)
-                ubyte sy = toscreeny(rotatedy[i], rz)
+            float rz = rotatedz[i]
+            if rz >= 0.1 {
+                ubyte sx = toscreenx(rotatedx[i], rz) as ubyte
+                ubyte sy = toscreeny(rotatedy[i], rz) as ubyte
                 c64scr.setchrclr(sx, sy, 46, i+2)
             }
         }
 
         for ubyte i in 0 to len(xcoor)-1 {
-            word rz = rotatedz[i]
-            if rz < 100 {
-                ubyte sx = toscreenx(rotatedx[i], rz)
-                ubyte sy = toscreeny(rotatedy[i], rz)
+            float rz = rotatedz[i]
+            if rz < 0.1 {
+                ubyte sx = toscreenx(rotatedx[i], rz) as ubyte
+                ubyte sy = toscreeny(rotatedy[i], rz) as ubyte
                 c64scr.setchrclr(sx, sy, 81, i+2)
             }
         }
diff --git a/compiler/examples/test.p8 b/compiler/examples/test.p8
index 27caadf28..9e9191de1 100644
--- a/compiler/examples/test.p8
+++ b/compiler/examples/test.p8
@@ -2,44 +2,15 @@
 
 ~ main {
 
-    sub start() {
-        memory byte b1 = $c000
-        memory byte b2 = $c001
-        memory word w1 = $c002
-        memory word w2 = $c004
-
-        float x =4.34
-        ubyte xx= x as ubyte
-        float y = x * 5.55
-        y =xx as float
-
-        b1=50
-        b2=-50
-        w1=100
-        w2=-100
-
-        c64scr.print_b(b1)
-        c64.CHROUT('/')
-        lsr(b1)
-        c64scr.print_b(b1)
-        c64.CHROUT('\n')
-        c64scr.print_b(b2)
-        c64.CHROUT('/')
-        lsr(b2)
-        c64scr.print_b(b2)
-        c64.CHROUT('\n')
-        c64scr.print_w(w1)
-        c64.CHROUT('/')
-        lsr(w1)
-        c64scr.print_w(w1)
-        c64.CHROUT('\n')
-        c64scr.print_w(w2)
-        c64.CHROUT('/')
-        lsr(w2)
-        c64scr.print_w(w2)
-        c64.CHROUT('\n')
-
-
+    word[1] rotatedx
 
+    sub start()  {
+        word xc = 2
+rpt:
+        word w = 2*xc
+        rotatedx[0] = w       ; @ok!
+        rotatedx[0] = 2*xc       ; @todo wrong code generated? crash!
+        c64.CHROUT('.')
+        goto rpt
     }
 }
diff --git a/compiler/src/prog8/compiler/target/c64/AsmGen.kt b/compiler/src/prog8/compiler/target/c64/AsmGen.kt
index 3a1dcd389..457c02f0a 100644
--- a/compiler/src/prog8/compiler/target/c64/AsmGen.kt
+++ b/compiler/src/prog8/compiler/target/c64/AsmGen.kt
@@ -564,7 +564,6 @@ class AsmGen(val options: CompilationOptions, val program: IntermediateProgram,
             Opcode.WRITE_INDEXED_VAR_WORD -> {
                 """
                 inx
-                inx
                 lda  ${ESTACK_LO.toHex()},x
                 asl  a
                 tay
diff --git a/prog8lib/prog8lib.p8 b/prog8lib/prog8lib.p8
index 0d8042cd2..d97345f6b 100644
--- a/prog8lib/prog8lib.p8
+++ b/prog8lib/prog8lib.p8
@@ -1006,7 +1006,7 @@ mult16		lda  #$00
 		sta  multiply_words_result+2	; clear upper bits of product
 		sta  multiply_words_result+3
 		ldx  #16			; for all 16 bits...
--	 	lsr  c64.SCRATCH_ZPWORD1+1		; divide multiplier by 2
+-	 	lsr  c64.SCRATCH_ZPWORD1+1	; divide multiplier by 2
 		ror  c64.SCRATCH_ZPWORD1
 		bcc  +
 		lda  multiply_words_result+2	; get upper half of product and add multiplicand