diff --git a/gr-sim/tfv_flying_6502.c b/gr-sim/tfv_flying_6502.c
index edeea155..c27fe8a7 100644
--- a/gr-sim/tfv_flying_6502.c
+++ b/gr-sim/tfv_flying_6502.c
@@ -83,7 +83,7 @@ static unsigned char water_map[32]={
 #define LOWRES_W	40
 #define LOWRES_H	40
 
-static int displayed=0;
+static int displayed=1;
 
 struct cycle_counts {
 	int flying;
@@ -229,18 +229,11 @@ static void fixed_mul(unsigned char x_i, unsigned char x_f,
 	num1h=x_i;
 	num1l=x_f;
 
-	negate=0;				// lda	#0		2
-						// sta	NEGATE		4
-
-						// lda	NUM1H		4
-
-						cycles.multiply+=13;
-
-
-	if (!(num1h&0x80)) goto check_num2;	// bpl check_num2	2nt/3
+	negate=0;
+						cycles.multiply+=10;
+	if (!(num1h&0x80)) goto check_num2;
 						cycles.multiply--;
-
-	negate++;				// inc NEGATE		6
+	negate++;
 
 	num1l=~num1l;
 	num1h=~num1h;
@@ -254,13 +247,12 @@ static void fixed_mul(unsigned char x_i, unsigned char x_f,
 
 	num1l&=0xff;
 	num1h&=0xff;
-						// total=26
-						cycles.multiply+=26;
+						cycles.multiply+=25;
 check_num2:
 
 	num2h=y_i;
 	num2l=y_f;
-						cycles.multiply+=7;
+						cycles.multiply+=6;
 	if (!(num2h&0x80)) goto unsigned_multiply;
 						cycles.multiply--;
 	negate++;
@@ -277,7 +269,7 @@ check_num2:
 
 	num2l&=0xff;
 	num2h&=0xff;
-						cycles.multiply+=30;
+						cycles.multiply+=25;
 unsigned_multiply:
 
 //	if (debug) {
@@ -290,7 +282,7 @@ unsigned_multiply:
 	aa=0;		// lda #0 (sz)
 	result2=aa;	// sta result+2
 	xx=16;		// ldx #16 (sz)
-						cycles.multiply+=8;
+						cycles.multiply+=7;
 multiply_mainloop:
 	cc=(num2h&1);	//lsr NUM2+1 (szc)
 	num2h>>=1;
@@ -306,7 +298,7 @@ multiply_mainloop:
 
 	num2l|=(cc<<7);
 	cc=cc2;
-						cycles.multiply+=15;
+						cycles.multiply+=13;
 	if (cc==0) goto shift_output;	// bcc L2
 						cycles.multiply--;
 
@@ -321,7 +313,7 @@ multiply_mainloop:
 	aa=aa+cc+num1h;			// adc NUM1+1
 	cc=!!(aa&0x100);
 	aa=aa&0xff;
-						cycles.multiply+=22;
+						cycles.multiply+=18;
 shift_output:
 	cc2=aa&1;
 	aa=aa>>1;
@@ -348,7 +340,7 @@ shift_output:
 	cc=cc2;		// ror result+0
 
 	xx--;				// dex
-						cycles.multiply+=25;
+						cycles.multiply+=22;
 	if (xx!=0) goto multiply_mainloop;	// bne L1
 						cycles.multiply--;
 	result3=aa&0xff;		// sta result+3
@@ -358,7 +350,7 @@ shift_output:
 //		printf("RAW RESULT = %02x:%02x:%02x:%02x\n",
 //			result3&0xff,result2&0xff,result1&0xff,result0&0xff);
 //	}
-						cycles.multiply+=13;
+						cycles.multiply+=11;
 	if (negate&1) {
 //		printf("NEGATING!\n");
 						cycles.multiply--;
@@ -383,7 +375,7 @@ shift_output:
 		aa-=result3+cc;
 		cc=!!(aa&0x100);
 		result3=aa;
-							cycles.multiply+=50;
+							cycles.multiply+=42;
 	}
 
 	*z_i=result2&0xff;
@@ -619,6 +611,21 @@ int flying(void) {
 	/* Flying					*/
 	/************************************************/
 
+	/* Benchmark the multiply */
+	memset(&cycles,0,sizeof(cycles));
+	fixed_mul(0x1,0x0,
+		0x2,0x0,
+		&ram[FACTOR_I],&ram[FACTOR_F]);
+	printf("Multiplying 1.0 * 2.0 = %d.%d, took %d cycles\n",
+		ram[FACTOR_I],ram[FACTOR_F],cycles.multiply);
+
+	memset(&cycles,0,sizeof(cycles));
+	fixed_mul(0xff,0xff,
+		0xff,0xff,
+		&ram[FACTOR_I],&ram[FACTOR_F]);
+	printf("Multiplying ff.ff * ff.ff = %d.%d, took %d cycles\n",
+		ram[FACTOR_I],ram[FACTOR_F],cycles.multiply);
+
 	gr();
 	clear_bottom(PAGE0);
 	clear_bottom(PAGE1);
diff --git a/tfv/OPTIMIZATION b/tfv/OPTIMIZATION
new file mode 100644
index 00000000..f0c31bdc
--- /dev/null
+++ b/tfv/OPTIMIZATION
@@ -0,0 +1,32 @@
+Original implementation:
+	Multiplying 1.0 * 2.0 = 2.0, took 707 cycles
+	Multiplying ff.ff * ff.ff = 0.0, took 761 cycles
+
+	Cycles: flying=               162
+	Cycles: getkey=                46
+	Cycles: page_flip=             26
+	Cycles: multiply=          88,179
+	Cycles: mode7=             76,077
+	Cycles: lookup_map=        33,920
+	Cycles: put_sprite=         2,561
+	==================================
+	Total =                   200,971
+	Frame Rate = 4.98 fps
+
+Update Multiply to use zero page addresses:
+	Multiplying 1.0 * 2.0 = 2.0, took 616 cycles
+	Multiplying ff.ff * ff.ff = 0.0, took 664 cycles
+
+	Cycles: flying=              162
+	Cycles: getkey=               46
+	Cycles: page_flip=            26
+	Cycles: multiply=         76,561
+	Cycles: mode7=            76,077
+	Cycles: lookup_map=       33,920
+	Cycles: put_sprite=        2,561
+	===================================
+	Total =                  189,353
+	Frame Rate = 5.28 fps
+
+
+
diff --git a/tfv/tfv_multiply.s b/tfv/tfv_multiply.s
index 10e80a36..c1801ab3 100644
--- a/tfv/tfv_multiply.s
+++ b/tfv/tfv_multiply.s
@@ -3,12 +3,12 @@
 ; NUM2 is zeroed out
 ; result is in RESULT3:RESULT2:RESULT1:RESULT0
 
-NUM1L:	.byte 0
-NUM1H:	.byte 0
-NUM2L:	.byte 0
-NUM2H:	.byte 0
-RESULT:	.byte 0,0,0,0
-NEGATE:	.byte 0
+;NUM1L:	.byte 0
+;NUM1H:	.byte 0
+;NUM2L:	.byte 0
+;NUM2H:	.byte 0
+;RESULT:	.byte 0,0,0,0
+;NEGATE:	.byte 0
 
 ; If we have 2k to spare we should check out
 ; http://codebase64.org/doku.php?id=base:seriously_fast_multiplication
@@ -16,94 +16,113 @@ NEGATE:	.byte 0
 multiply:
 
 	lda	#$0							; 2
-	sta	NEGATE							; 4
+	sta	NEGATE							; 3
 
 	; Handle Signed
-	lda	NUM1H							; 4
+	lda	NUM1H							; 3
 	bpl	check_num2						; 2nt/3
+								;==============
+								;	 10
 
-	inc	NEGATE							; 4
+	inc	NEGATE							; 3
 
 	clc		; 2s-complement NUM1H/NUM1L			; 2
-	lda	NUM1L							; 4
+	lda	NUM1L							; 3
 	eor	#$ff							; 2
 	adc	#$1							; 2
-	sta	NUM1L							; 4
+	sta	NUM1L							; 3
 
-	lda	NUM1H							; 4
+	lda	NUM1H							; 3
 	eor	#$ff							; 2
 	adc	#$0							; 2
-	sta	NUM1H							; 4
-
+	sta	NUM1H							; 3
+								;===========
+								;	 25
 check_num2:
-	lda	NUM2H							; 4
+	lda	NUM2H							; 3
 	bpl	unsigned_multiply					; 2nt/3
+								;==============
+								;	  6
 
-	inc	NEGATE							; 4
+	inc	NEGATE							; 3
 
 	clc								; 2
-	lda	NUM2L							; 4
+	lda	NUM2L							; 3
 	eor	#$ff							; 2
 	adc	#$1							; 2
-	sta	NUM2L							; 4
+	sta	NUM2L							; 3
 
-	lda	NUM2H							; 4
+	lda	NUM2H							; 3
 	eor	#$ff							; 2
 	adc	#$0							; 2
-	sta	NUM2H							; 4
-
+	sta	NUM2H							; 3
+								;=============
+								;	 25
 unsigned_multiply:
 
 	lda	#0		; Initialize RESULT to 0		; 2
-	sta 	RESULT+2						; 4
+	sta 	RESULT+2						; 3
 	ldx	#16		; 16x16 multiply			; 2
+								;============
+								;	  7
 multiply_mainloop:
-	lsr	NUM2H		; Shift right 16-bit NUM2		; 6
-	ror	NUM2L		; low bit goes into carry		; 6
+	lsr	NUM2H		; Shift right 16-bit NUM2		; 5
+	ror	NUM2L		; low bit goes into carry		; 5
 	bcc	shift_output	; 0 or 1?				; 2nt/3
+								;============
+								;	 13
+
 	tay			; If 1, add NUM1 (hi byte RESULT in A)	; 2
 	clc								; 2
-	lda	NUM1L							; 4
-	adc	RESULT+2						; 4
-	sta	RESULT+2						; 4
+	lda	NUM1L							; 3
+	adc	RESULT+2						; 3
+	sta	RESULT+2						; 3
 	tya								; 2
-	adc	NUM1H							; 4
+	adc	NUM1H							; 3
+								;============
+								;	 18
 shift_output:
 	ror	A		; "Stairstep" shift			; 2
-	ror	RESULT+2						; 6
-	ror	RESULT+1						; 6
-	ror	RESULT							; 6
+	ror	RESULT+2						; 5
+	ror	RESULT+1						; 5
+	ror	RESULT							; 5
 	dex								; 2
 	bne	multiply_mainloop					; 2nt/3
-	sta	RESULT+3						; 4
+								;=============
+								;	 22
+
+	sta	RESULT+3						; 3
 
 	;; Negate if necessary
 
-	lda	NEGATE							; 4
+	lda	NEGATE							; 3
 	and	#$1							; 2
 	beq	positive						; 2nt/3
+								;==============
+								;	 11
 
 	clc								; 2
-	lda	RESULT+0						; 4
+	lda	RESULT+0						; 3
 	eor	#$ff							; 2
 	adc	#$1							; 2
-	sta	RESULT+0						; 4
+	sta	RESULT+0						; 3
 
-	lda	RESULT+1						; 4
+	lda	RESULT+1						; 3
 	eor	#$ff							; 2
 	adc	#$0							; 2
-	sta	RESULT+1						; 4
+	sta	RESULT+1						; 3
 
-	lda	RESULT+2						; 4
+	lda	RESULT+2						; 3
 	eor	#$ff							; 2
 	adc	#$0							; 2
-	sta	RESULT+2						; 4
+	sta	RESULT+2						; 3
 
-	lda	RESULT+3						; 4
+	lda	RESULT+3						; 3
 	eor	#$ff							; 2
 	adc	#$0							; 2
-	sta	RESULT+3						; 4
-
+	sta	RESULT+3						; 3
+								;===========
+								;	 42
 positive:
 
 	rts								; 6
diff --git a/tfv/tfv_zp.inc b/tfv/tfv_zp.inc
index 592b1f78..df8f6d0e 100644
--- a/tfv/tfv_zp.inc
+++ b/tfv/tfv_zp.inc
@@ -1,6 +1,7 @@
 .define EQU =
 
 ;; Zero page monitor routines addresses
+
 WNDLFT	EQU $20
 WNDWDTH	EQU $21
 WNDTOP	EQU $22
@@ -17,6 +18,40 @@ MASK	EQU $2E
 COLOR	EQU $30
 INVFLG	EQU $32
 
+; More zero-page addresses
+; we try not to conflict with anything DOS, MONITOR or BASIC related
+
+COLOR1		EQU	$E0
+COLOR2		EQU	$E1
+MATCH		EQU	$E2
+XX		EQU	$E3
+YY		EQU	$E4
+YADD		EQU	$E5
+LOOP		EQU	$E6
+MEMPTRL		EQU	$E7
+MEMPTRH		EQU	$E8
+NAMEL		EQU	$E9
+NAMEH		EQU	$EA
+NAMEX		EQU	$EB
+CHAR		EQU	$EC
+DISP_PAGE       EQU	$ED
+DRAW_PAGE       EQU	$EE
+
+FIRST		EQU	$F0
+LASTKEY		EQU	$F1
+PADDLE_STATUS	EQU	$F2
+XPOS		EQU	$F3
+YPOS		EQU	$F4
+TEMP		EQU	$FA
+RUN		EQU	$FA
+TEMP2		EQU	$FB
+TEMPY		EQU	$FB
+INL		EQU	$FC
+INH		EQU	$FD
+OUTL		EQU	$FE
+OUTH		EQU	$FF
+
+
 ;; Flying Routine Only
 
 TURNING         EQU     $60
@@ -49,6 +84,13 @@ DRAW_SPLASH     EQU     $7A
 SPEED           EQU     $7B
 SPLASH_COUNT    EQU     $7C
 OVER_WATER      EQU     $7D
+NUM1L		EQU	$7E
+NUM1H		EQU	$7F
+NUM2L		EQU	$80
+NUM2H		EQU	$81
+RESULT		EQU	$82 ; 83,84,85
+NEGATE		EQU	$86
+
 SHIPY           EQU     $E4
 
 ;; World Map Only
@@ -110,38 +152,6 @@ COUT1	EQU $FDF0			;; output A to screen
 
 
 
-; Our zero-page addresses
-; we try not to conflict with anything DOS, MONITOR or BASIC related
-
-COLOR1		EQU	$E0
-COLOR2		EQU	$E1
-MATCH		EQU	$E2
-XX		EQU	$E3
-YY		EQU	$E4
-YADD		EQU	$E5
-LOOP		EQU	$E6
-MEMPTRL		EQU	$E7
-MEMPTRH		EQU	$E8
-NAMEL		EQU	$E9
-NAMEH		EQU	$EA
-NAMEX		EQU	$EB
-CHAR		EQU	$EC
-DISP_PAGE       EQU	$ED
-DRAW_PAGE       EQU	$EE
-
-FIRST		EQU	$F0
-LASTKEY		EQU	$F1
-PADDLE_STATUS	EQU	$F2
-XPOS		EQU	$F3
-YPOS		EQU	$F4
-TEMP		EQU	$FA
-RUN		EQU	$FA
-TEMP2		EQU	$FB
-TEMPY		EQU	$FB
-INL		EQU	$FC
-INH		EQU	$FD
-OUTL		EQU	$FE
-OUTH		EQU	$FF