X-Git-Url: https://git.gir.st/Chiptunes.git/blobdiff_plain/8ee3310ee20cbcc5b69824e1ab77a8fb34b147f9..44e34da036f83d3efd6384cdc37ec9bd3e234e3a:/foo.c

diff --git a/foo.c b/foo.c
index d708f3c..c2bcb5d 100644
--- a/foo.c
+++ b/foo.c
@@ -6,14 +6,22 @@ u8 data[] = {
 	0x84, 0x9d, 0xb0, 0x69, 0x9d, 0x84, 0x69, 0x58,
 	0x75, 0x8c, 0xb0, 0x69, 0x8c, 0x75, 0x69, 0x58
 };
-u8 zero; //zero register
-u8 i0;
-u8 i1;
-u8 i2;
-u8 i3;
-u8 t;
-u8 x;
-u8 _;
+u8 zero;	//r16
+u8 acc;		//r17
+u8 i0;		//r18
+u8 i1;		//r19
+u8 i2;		//r20
+u8 i3;		//r21
+u8 n;		//r22
+u8 s;		//r23
+u8 _;		//r24
+u8 loop;	//r25
+u8 t;/*==Ml*/	//r26 (Xlo)
+u8 x;/*==Mh*/	//r27 (Xhi)
+		//r28
+		//r29
+/*fakestack_l*/	//r30 (Zlo)
+/*fakestack_h*/	//r31 (Zhi)
 #define Mh x //mod3 vars
 #define Ml t // -"-
 //http://homepage.divms.uiowa.edu/~jones/bcd/mod.shtml
@@ -47,6 +55,219 @@ void mod3(void) {
 	RET
 	#undef tmp
 }
+void mul(void) { //don't need overhead of function (inline it)
+	// i1.i0 * t -> _.x.t
+	#define a1 x
+	#define a2 _
+	#define a0 t
+	// start MUL -- 92 cycles :( (unrolled and skipping second bit: 76)
+	CLR	(a2)
+	CLR	(a1)
+
+	CPI	(t, 0x58)
+	BREQ	(mul_58)
+	CPI	(t, 0x69)
+	BREQ	(mul_69)
+	CPI	(t, 0x75)
+	BREQ	(mul_75)
+	CPI	(t, 0x84)
+	BREQ	(mul_84)
+	CPI	(t, 0x8c)
+	BREQ	(mul_8c)
+	CPI	(t, 0x9d)
+	BREQ	(mul_9d)
+	CPI	(t, 0xb0)
+	BREQ	(mul_b0)
+	mul_58: // 0101 1000
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_69: // 0110 1001
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_75: // 0111 0101
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_84: // 1000 0100
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_8c: // 1000 1100
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_9d: // 1001 1101
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		RJMP	(endmul)
+	mul_b0: // 1011 0000
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+		LSR (a2)
+		ROR (a1)
+		ADD (a1, i0)
+		ADC (a2, i1, carry)
+		LSR (a2)
+		ROR (a1)
+	endmul:
+
+	// end MUL
+	#undef a0
+	#undef a1
+	#undef a2
+	RET
+}
 void g(void) {
 	// g(i, t) -> t
 	// tempvars: `x` and `_`
@@ -65,15 +286,17 @@ void g(void) {
 	  ADD X_lo, t
 	  ADC X_hi, zero
 	  LD  t, X         */
-	t = (((i1&0x1f)<<8|i0)*t)>>8; //TODO
-	RET //TODO: CALL/RET is expensive; store next-instruction-position in Z register and RJMP, then JRMP back (maybe unnecessary, since timer might need stack anyways)
+	RCALL	mul(); //stores used value in in x
+	MOV	(t, x)
+	RET //TODO: replace CALL/RET with IJMP?
 };
 
 int main(void) {
-	u8 n;
-	u8 s;
-	u8 acc;
-	//TODO: clear all vars/registers
+	CLR	(zero)
+	CLR	(i0)
+	CLR	(i1)
+	CLR	(i2)
+	CLR	(i3)
 	for (;;) {
 		MOV	(n, i2)
 		LSL	(n)
@@ -87,7 +310,7 @@ int main(void) {
 		OR	(n, tmp)
 		#undef tmp
 		MOV	(s, i3)
-		ROR	(s)
+		LSR	(s)
 		ROR	(s)
 		ANDI	(s, 0x80)
 		#define tmp _
@@ -100,7 +323,6 @@ int main(void) {
 		MOV	(t, n)
 		RCALL	g();
 		SWAP	(t)
-		ANDI	(t, 0x0f)
 		ANDI	(t, 1)
 		MOV	(acc, t)
 
@@ -168,9 +390,8 @@ int main(void) {
 		ADD	(tmp, x)
 		ROR	(tmp)
 		LSR	(tmp)
-                MOV	(x, tmp)
+		AND	(t, tmp)
 		#undef tmp
-		AND	(t, x)
 		ADD	(acc, t)
 
 		//voice 4:
@@ -203,24 +424,22 @@ int main(void) {
 		RCALL	g();
 		LSR	(t)
 		ANDI	(t, 3)
-		MOV	(x, s)
-		INC	(x)
+		INC	(s)
 		#define tmp _
-                MOV	(tmp, x)
+                MOV	(tmp, s)
 		LSR	(tmp)
-                ADD	(tmp, x)
+                ADD	(tmp, s)
 		ROR	(tmp)
 		LSR	(tmp)
 		LSR	(tmp)
-                ADD	(tmp, x)
+                ADD	(tmp, s)
 		ROR	(tmp)
-                ADD	(tmp, x)
+                ADD	(tmp, s)
 		ROR	(tmp)
 		LSR	(tmp)
 		LSR	(tmp)
-                MOV	(x, tmp)
+		AND	(t, tmp)
 		#undef tmp
-		AND	(t, x)
 		ADD	(acc, t)
 
 		putchar(acc<<4); //TODO