ALU 2, @12, KC0[], KC1[]
	TEX 2 @6
	ALU 40, @15, KC0[CB0:0-32], KC1[]
	MEM_RAT_CACHELESS STORE_RAW T0.X, T1.X, 1
	CF_END
	PAD

	Fetch clause starting at 6:
	  VTX_READ_32 T4.X, T4.X, 44, #3 ; goffsetX
	  VTX_READ_32 T3.X, T3.X, 48, #3 ; goffsetY
	  VTX_READ_32 T2.X, T2.X, 52, #3 ; goffsetZ

	ALU clause starting at 12:
	  MOV * T2.X, 0.0,  
	  MOV   T3.X, PV.X,  
	  MOV * T4.X, PV.X,  

	ALU clause starting at 15:
	  MULLO_INT * T0.W, KC0[2].X, T1.Z,  ; T0.W <-- LSIZE.Z * WGID.Z
	  ADD_INT   T0.W, PS, T0.Z,          ; T0.W <-- T0.W + LID.Z
	  INT_TO_FLT * T0.Z, KC0[0].W,       ; T0.Z <-- flt(GSIZE.X)
	  ADD_INT   T0.W, PV.W, T2.X,        ; T0.W <-- T0.W + GOFFSET.Z ;; T0.W == gid.z
	  RECIP_IEEE * T1.Z, PS,             ; T1.Z <-- 1/T0.Z ;; 1/flt(gsize.x)
	  INT_TO_FLT * T1.W, PV.W,           ; T1.W <-- flt(T0.W) ;; flt(gid.z)
	  INT_TO_FLT * T2.X, KC0[1].Y,       ; T2.X <-- flt(GSIZE.Z)  
	  INT_TO_FLT * T2.Y, KC0[1].X,       ; T2.Y <-- flt(GSIZE.Y)
	  MULLO_INT * T1.Y, KC0[1].W, T1.Y,  ; T1.Y <-- LSIZE.Y * WGID.Y
	  ADD_INT   T2.W, PS, T0.Y,          ; T2.W <-- T1.Y + LID.Y
	  MULLO_INT * T0.Y, KC0[1].Z, T1.X,  ; T0.Y <-- LSIZE.X * WGID.X
	  ADD_INT   T2.Z, PV.W, T3.X,        ; T2.Z <-- T2.W + GOFFSET.Y ;; T2.Z == gid.y
	  ADD_INT   T2.W, PS, T0.X,  BS:VEC_021/SCL_122 ; T2.W <-- T0.Y + LID.X
	  RECIP_IEEE * T0.X, T2.Y,           ; T0.X <-- 1/T2.Y ;; 1/flt(gsize.y)
	  ADD_INT   T2.W, PV.W, T4.X,        ; T2.W <-- T2.W + GOFFSET.X ;; T2.W == gid.x
	  INT_TO_FLT * T0.Y, PV.Z,           ; T0.Y <-- flt(T2.Z) ;; flt(gid.y)
	  MULADD_IEEE   T3.W, T2.Y, literal.x, PS, ; T3.W <-- T2.Y * -0.5 + T0.Y ;; flt(gid.y) - flt(gsize.y)/2
	  INT_TO_FLT * T0.Y, PV.W,           ; T0.Y <-- flt(T2.W) ;; flt(gid.x)
	-1090519040(-5.000000e-01), 0(0.000000e+00)
	  MULADD_IEEE   T0.Z, T0.Z, literal.x, PS, ; T0.Z <-- T0.Z * -0.5 + T0.Y ;; flt(gid.x) - flt(gsize.x)/2
	  MUL_IEEE   T3.W, PV.W, T0.X,             ; T3.W <-- T3.W * T0.X ;; (flt(gid.y) - flt(gsize.y)/2) * 1/flt(gsize.y)
	  MULLO_INT * T0.X, T2.W, KC0[1].X,        ; T0.X <-- T2.W * gsize.y ;; gid.x * gsize.y
	-1090519040(-5.000000e-01), 0(0.000000e+00)
	  ADD_INT   T0.X, PS, T2.Z,                  ; T0.X <-- T0.X + gid.y
	  MULADD_IEEE   T0.Y, T2.X, literal.x, T1.W, ; T0.Y <-- T2.X * -0.5 + T1.W ;; flt(gid.z) - flt(gsize.Z)/2
	  MUL_IEEE   T2.Z, PV.W, PV.W,               ; T2.Z <-- T3.W * T3.W
	  MUL_IEEE   T1.W, PV.Z, T1.Z,  BS:VEC_021/SCL_122 ; T1.W <-- T0.Z * T1.Z ;; (flt(gid.x) - flt(gsize.x)/2) * 1/flt(gsize.x)
	  RECIP_IEEE * T0.Z, T2.X,                   ; T0.Z <-- 1/T2.X ;; 1/flt(gsize.z)
	-1090519040(-5.000000e-01), 0(0.000000e+00)
	  MULADD_IEEE   T1.Z, PV.W, PV.W, PV.Z,      ; T1.Z <-- T1.W * T1.W + T2.Z ;; xf^2 + yf^2
	  MUL_IEEE   T1.W, PV.Y, PS,                 ; T1.W <-- T0.Y * T0.Z ;; (flt(gid.z) - flt(gsize.Z)/2) * 1/flt(gsize.z)
	  MULLO_INT * T0.X, PV.X, KC0[1].Y,          ; T0.X <-- T0.X * gsize.z
	  MULADD_IEEE   T1.W, PV.W, PV.W, PV.Z,      ; T1.W <-- T1.W * T1.W + T1.Z ;; zf^2 + xf^2 + yf^2
	  ADD_INT * T0.W, PS, T0.W,                  ; T0.W <-- T0.X + gid.x
	  LSHL   T0.W, PS, literal.x,                ; T0.W <-- T0.W << 2 ;; int index to ptr offset
	  SETGT_DX10 * T1.W, literal.y, PV.W,        ; T1.W <-- 0.25 > T1.W ? 0xfffffff : 0x0
	2(2.802597e-45), 1048576000(2.500000e-01)
	  AND_INT   T0.X, PS, 1,                     ; T0.X <-- T1.W & 0x1
	  ADD_INT * T0.W, KC0[2].Y, PV.W,            ; T0.W <-- out + T0.W
	  LSHR * T1.X, PV.W, literal.x,              ; T1.X <-- T0.W >> 2 ;; mem_rat operates on dword addresses
	2(2.802597e-45), 0(0.000000e+00)