Differences

This shows you the differences between two versions of the page.

--- base:fastest_multiplication [2017-04-17 13:33] – Added heading and credits at start of the article for standardization across the wiki. ftc
+++ base:fastest_multiplication [2017-04-19 09:47] – repose
@@ Line 7: / Line 7: @@
 Jack Asser's: 233 cycles  ref: [[seriously_fast_multiplication]] \\
 Chris Jam's: 204.5  ref: http://csdb.dk/forums/?roomid=11&topicid=91766 \\
-Mine: 201 zp variation: 198 \\
+Mine: 196 zp variation: 192 \\
 Times above need to add 12 for jsr/rts \\
@@ Line 40: / Line 40: @@
 z3=$83
-;not shown is a routine to make the tables
+;Example showing use
-;also you need to init the pointers' high bytes to the tables
+lda #$ff
+sta x0
+sta x1
+sta y0
+sta y1
+jsr maketables
+jsr umult16
+;result should be $fffe0001, e.g. as viewed with a typical m 0080 monitor command:
+;0080 01 00 fe ff
+makesqrtables:
+;init zp square tables pointers
+lda #>sqrlo
+sta p_sqr_lo+1
+lda #>sqrhi
+sta p_sqr_hi+1
+lda #>negsqrlo
+sta p_invsqr_lo+1
+lda #>negsqrhi
+sta p_invsqr_hi+1
+;generate sqr(x)=x^2/4
+      ldx #$00
+      txa
+      !by $c9   ; CMP #immediate - skip TYA and clear carry flag
+lb1:  tya
+      adc #$00
+ml1:  sta sqrhi,x
+      tay
+      cmp #$40
+      txa
+      ror
+ml9:  adc #$00
+      sta ml9+1
+      inx
+ml0:  sta sqrlo,x
+      bne lb1
+      inc ml0+2
+      inc ml1+2
+      clc
+      iny
+      bne lb1
+;generate negsqr(x)=(255-x)^2/4
+      ldx #$00
+      ldy #$ff
+mt1:
+      lda sqrhi+1,x
+      sta negsqrhi+$100,x
+      lda sqrhi,x
+      sta negsqrhi,y
+      lda sqrlo+1,x
+      sta negsqrlo+$100,x
+      lda sqrlo,x
+      sta negsqrlo,y
+      dey
+      inx
+      bne mt1
+      rts
 umult16:
@@ Line 95: / Line 153: @@
 lda (p_sqr_hi),y
 sbc (p_invsqr_hi),y
-sta z3;x1*y1h;31
+tay;x1*y1h;Y=z3, 30 cycles
+;17+33+31+17+31+30=159
-;4*31+2*17 so far=158
-;add partials
-;-add first two numbers in column 1
-;jmp do_adds;put in zp to save 3 cycles :)
 do_adds:
-clc
+;-add the first two numbers of column 1
-c1a lda #0
+	clc
-c1b adc #0
+c1a:	lda #0
-sta z1;9
+c1b:	adc #0
-;-continue to first two numbers in column 2
+	sta z1;9
-c2a lda #0
-c2b adc #0
+;-continue to first two numbers of column 2
-sta z2;7
+c2a:	lda #0
-bcc c1c;3 taken/9 not taken, avg 6
+c2b:	adc #0
-inc z3
+	tax;X=z2, 6 cycles
-clc
+	bcc c1c;3/6 avg 4.5
-;-add last number of column 1 (row c)
+	iny;z3++
-c1c lda #0
+	clc
-adc z1
-sta z1;8
+;-add last number of column 1
+c1c:	lda #0
+	adc z1
+	sta z1;8
 ;-add last number of column 2
-c2c lda #0
+	txa;A=z2
-adc z2
+c2c:	adc #0
-sta z2;8
+	tax;X=z2, 6
-bcc fin;3/7 avg 5
+	bcc fin;3/4 avg 3.5
-inc z3
+	iny;z3++
-;9+7+6+8+8+5=43
-fin rts
+;Y=z3, X=z2
+;9+6+4.5+8+6+3.5=37
+fin:	rts
 Diagram of the additions