optimise pixel draw using lookup table

wip-hugo/END.s

@@ -1,2 +1,55 @@
 binary_factor:
 .byte   %00000001, %00000010, %00000100, %00001000, %00010000, %00100000, %01000000, %10000000
+
+;;This is used by pixel_draw! look at it for more detail
+big_y_offset:
+        .lobytes $0000
+        .hibytes $0000
+        .lobytes  $0140
+        .hibytes $0140
+        .lobytes  $0280
+        .hibytes $0280
+        .lobytes  $03C0
+        .hibytes $03C0
+        .lobytes  $0500
+        .hibytes $0500
+        .lobytes  $0640
+        .hibytes $0640
+        .lobytes  $0780
+        .hibytes $0780
+        .lobytes  $08C0
+        .hibytes $08C0
+        .lobytes  $0A00
+        .hibytes $0A00
+        .lobytes  $0B40
+        .hibytes $0B40
+        .lobytes  $0C80
+        .hibytes $0C80
+        .lobytes  $0DC0
+        .hibytes $0DC0
+        .lobytes  $0F00
+        .hibytes $0F00
+        .lobytes  $1040
+        .hibytes $1040
+        .lobytes  $1180
+        .hibytes $1180
+        .lobytes  $12C0
+        .hibytes $12C0
+        .lobytes  $1400
+        .hibytes $1400
+        .lobytes  $1540
+        .hibytes $1540
+        .lobytes  $1680
+        .hibytes $1680
+        .lobytes  $17C0
+        .hibytes $17C0
+        .lobytes  $1900
+        .hibytes $1900
+        .lobytes  $1A40
+        .hibytes $1A40
+        .lobytes  $1B80
+        .hibytes $1B80
+        .lobytes  $1CC0
+        .hibytes $1CC0
+        .lobytes  $1E00
+        .hibytes $1E00

wip-hugo/routines/pixel/pixel.inc

@@ -1,9 +1,9 @@
 ;;; -*- Mode: asm; indent-tabs-mode: t; tab-width: 8 -*-
-      ;; public args
+;; Note that this file should never collide with line.inc
+        ;; public args
        Y_pos = ARGVEC + 0
        X_pos = ARGVEC + 1
        ;; Private args
        byte_to_paint = $E8
        btp_mem_pos = $E9  ; 16-bit value (uses EA), byte to paint memory position
-       C = $EB            ; 16-bit value (uses EC)
+       Bitmap = $4000
-       B = $ED            ; 16-bit value (uses EE)

wip-hugo/routines/pixel/pixel_draw.s

@@ -2,75 +2,66 @@
 
 ;;Screen print. Draws a pixel at a specified position.
 .proc pixel_draw; user-procedure :clobbers (A X Y) :clobbers-arguments 0
-;; Draws a pixel at Y_pos, X_pos se mem.inc
         .include "pixel.inc"
-        Bitmap = $4000
+
-        Bitmap_end = $5F3F
+        ;; Lets clear some memory for later
+        LDA #$00
+	STA btp_mem_pos + 1
+
+calc_byte_to_paint:
         ;; X = X_pos (mod 8)
-	LDA X_pos ; X (mod 8)
+	LDA X_pos
 	AND #%00000111
 	TAX
 
-	;;this is the same as: A = 2^X
+	;;this is the same as: byte_to_paint = 2^X
-	LDA binary_factor, X
+	LDA binary_factor, X;; (see END.s)
-
 	STA byte_to_paint
 
-	;;FIND THE POSITION IN MEMORY TO WRITE PIXEL
+	;;FIND position to write byte_to_paint (btp_mem_pos)
 	;; + + + + + >  X
 	;; +
 	;; +
 	;;\/
 	;; Y
-	;;
 
-	;;pos = x_offset
+	;; The screen has a wierd memory to pixel maping!
-	LDA #%11111000
+	;; Each line of 8 pixels is one byte
-	AND X_pos
+	;; 8 consecutive bytes in memory makes up a chunk and are 'under each other' on the screen
+	;; the chunks are orderd from the left to the right and then looping downwards.
+	;; Therefore we have that: btp_mem_pos = big_x_offset + smal_y_offset + big_y_offset    + bitmap_offset
+
+	;; We use a lookup-table for big_y_offset (see END.s)
+	;; (Y_pos / 8  floor)*2
+	LDA Y_pos
+	LSR A
+	LSR A
+	;; We need to remove the last digit if its still there
+	LSR A
+	ASL A
+
+	TAX
+	LDA big_y_offset, X
 	STA btp_mem_pos
-
+	INX
-	LDA #$00
+	LDA big_y_offset,X
 	STA btp_mem_pos + 1
 
-	;;The y_pos adds offset because chunk offsets + inside chunk offset.
+	;; btp_mem_pos = big_x_offset
-	;; Adding inside chunk offset
+	LDA #%11111000
+	AND X_pos
+	Add_16 btp_mem_pos, btp_mem_pos +1, A, #$00
+
+	;; + smal_y_offset
 	LDA #%00000111 ; A = y (mod 8)
 	AND Y_pos ;; offset to add
 	TAY
 
-	LDA #$00
-	STA $4B
-	;;y =8 translates to 320 bytes.
-	CLC
-	LDA #%11111000 ; A =  y - [y (mod 8)]
-	AND Y_pos
-	STA C
-	STA B
-	LDA #$00
-	STA C+1
-	STA B+1
-
-	;;We need to calculate C*40. 40 = 2*2*2*(2^2 +1)
-	;; _*2^2
-	Mult_16 C, C+1
-	Mult_16 C, C+1
-
-	;; + _*1
-	Add_16 C, C+1, B, B+1, !
-
-	;; *2*2*2
-	Mult_16 C, C + 1
-	Mult_16 C, C + 1
-	Mult_16 C, C + 1
-
-	Add_16 btp_mem_pos, btp_mem_pos + 1, C, C + 1, !
-
-	;;add offset for where bitmap is
 	Add_16 btp_mem_pos, btp_mem_pos + 1, #<Bitmap, #>Bitmap, !
 
 	;;Let draw some stuff
 	LDA byte_to_paint ;; note that both bytes are used!
-	ORA (>btp_mem_pos), Y
+	ORA (btp_mem_pos), Y
-	STA (>btp_mem_pos), Y
+	STA (btp_mem_pos), Y
 	RTS
 .endproc

wip-hugo/routines/pixel/pixel_test.s

@@ -0,0 +1,8 @@
+.scope pixel_test
+        .include "pixel.inc"
+        LDA #$40
+        STA X_pos
+        LDA #$40
+        STA Y_pos
+        jsr pixel_draw
+.endscope

wip-hugo/routines/text/char.inc

@@ -3,8 +3,6 @@
        Y_pos   = ARGVEC + 1
        code    = ARGVEC + 2
 
-       ;; private variables
-       screen_position = $EE ; 16-bit value (uses EF)
        ;; 16-bit value, we reuse bytes from X_pos
        X_pos_offset_hi = $ED
        X_pos_offset_lo = X_pos
@@ -12,11 +10,15 @@
        Y_pos_offset_hi = $EC
        Y_pos_offset_lo = Y_pos
 
+       ;; private variables
+       screen_position = $EE ; 16-bit value (uses EF)
+       temp_value = $EC      ; 16-bit value (uses ED)
+
        ;; What you will see bellow is .. interesting. I want to reuse
        ;; code and I need it for a 16-bit pointer. Because 6502 pointers expect the bytes to be besides each other
        ;; I need to use code +1 to store data even though its fot ARGVEK.
        ;; Thats why it's cursed
-       petski_position = code ; 16 bit value uses (code +1)
+       petski_position = code ; 16-bit value uses (code +1)
 
        ;; values
        VIC_bank = $4000

wip-hugo/source.s

@@ -7,8 +7,10 @@
 ;;Code to run
 .include "STARTUP.s"
 ;.include "dubbel_buffer/raster_irqs.s"
+
 ;.include "routines/line/line_test.s"
-.include "routines/text/char_draw_test.s"
+;.include "routines/text/char_draw_test.s"
+.include "routines/pixel/pixel_test.s"
 ;.include "routines/memory/memcpy_test.s"
 ;.include "routines/triangle/triangle_test.s"
 exit: