fixes

not yet doing any of the calc/output or calls :D

mandel.s

@@ -14,21 +14,44 @@ zx_zy = $94     ; fixed4.12: z_x * z_y
 dist  = $96     ; fixed4.12: z_x^2 + z_y^2
 
 iter         = $a0 ; u8: iteration count
-zoom  = $a1     ; u8: zoom shift level
-temp  = $a2     ; u16
-temp2 = $a4     ; u16
 
-pixel_ptr    = $b0 ; u16
-pixel_color  = $b2 ; u8
-pixel_mask   = $b3 ; u8
-pixel_shift  = $b4 ; u8
-pixel_offset = $b5 ; u8
+zoom         = $a1 ; u8: zoom shift level
+count_frames = $a2 ; u8
+count_pixels = $a3 ; u8
+total_ms     = $a4 ; float48
+total_pixels = $aa ; float48
+
+temp         = $b0 ; u16
+pixel_ptr    = $b2 ; u16
+pixel_color  = $b4 ; u8
+pixel_mask   = $b5 ; u8
+pixel_shift  = $b6 ; u8
+pixel_offset = $b7 ; u8
+
 
 ; FP registers in zero page
-FR0 = $d4
+FR0    = $d4 ; float48
 FRE    = $da
-FR1 = $e0
-FR2 = $e6
+FR1    = $e0 ; float48
+FR2    = $e6 ; float48
+CIX    = $f2 ; u8 - index into INBUFF
+INBUFF = $f3 ; u16 - pointer to ascii
+FLPTR  = $fc ; u16 - pointer to user buffer float48
+
+LBUFF  = $0580 ; result buffer for FASC routine
+
+; FP ROM routine vectors
+FASC   = $D8E6 ; FLOATING POINT TO ASCII (output in INBUFF, last char has high bit set)
+IFP    = $D9AA ; INTEGER TO FLOATING POINT CONVERSION (FR0:u16 -> FR0:float48)
+FADD   = $DA66 ; ADDITION       (FR0 += FR1)
+FSUB   = $DA60 ; SUBTRACTION    (FR0 -= FR1)
+FMUL   = $DADB ; MULTIPLICATION (FR0 *= FR1)
+FDIV   = $DB28 ; DIVISION       (FR0 /= FR1)
+ZF1    = $DA46 ; CLEAR ZERO PAGE FLOATING POINT NUMBER (XX)
+FLD0R  = $DD89 ; LOAD FR0 WITH FLOATING POINT NUMBER (YYXX)
+FLD1R  = $DD98 ; LOAD FR1 WITH FLOATING POINT NUMBER (YYXX)
+FST0R  = $DDA7 ; STORE FR0 IN USER BUFFER (YYXX)
+FMOVE  = $DDB6 ; MOVE FR0 TO FR1
 
 ; High data
 framebuffer_top    = $8000
@@ -51,6 +74,15 @@ DLISTH = $D403
 SDLSTL = $230
 SDLSTH = $231
 
+; interrupt stuff
+XITVBV = $E462
+SETVBV = $E45C
+
+.struct float48
+    exponent .byte
+    mantissa .byte 6
+.endstruct
+
 .data
 
 strings:
@@ -72,6 +104,10 @@ str_speed_len = str_speed_end - str_speed
 str_run_len = str_run_end - str_run
 str_done_len = str_done_end - str_done
 
+speed_start = str_self_len + 2
+speed_len = 14 + str_speed_len
+
+
 char_map:
     ; Map ATASCII string values to framebuffer font entries
     ; Sighhhhh
@@ -98,12 +134,19 @@ aspect:
     ;
     ; 184h is the equiv of 220.8h at square pixels
     ; 320 / 220.8 = 1.45 display aspect ratio
-aspect_x:
+aspect_x: ; fixed4.16 5/4
     .word 5 << (12 - 2)
 
-aspect_y:
+aspect_y: ; fixed4.16 3/4
     .word 3 << (12 - 2)
 
+ms_per_frame: ; float48 16.66666667
+    .byte 64  ; exponent/sign
+    .byte $16 ; BCD digits
+    .byte $66
+    .byte $66
+    .byte $66
+    .byte $67
 
 display_list_start:
     ; 24 lines overscan
@@ -216,6 +259,10 @@ color_map:
     copy 4, dest, arg
 .endmacro
 
+.macro copyfloat dest, arg
+    copy 6, dest, arg
+.endmacro
+
 ; 2 + 8 * byte cycles
 .macro neg bytes, arg
     sec ; 2 cyc
@@ -613,14 +660,38 @@ loop:
 done:
 .endmacro
 
+.proc vblank_handler
+    inc count_frames
+    jmp XITVBV
+.endproc
+
+.proc update_speed
+    ; convert frames (u16) to fp
+    ; add to frames_total
+    ; convert pixels (u16) to fp
+    ; add to pixels_total
+    ; (frames_total * 16.66666667) / pixels_total
+    ; convert to ATASCII
+    ; draw text
+.endproc
+
 .proc start
 
     ; ox = 0; oy = 0; zoom = 0
+    ; count_frames = 0; count_pixels = 0
     lda #0
     sta ox
     sta ox + 1
     sta oy
     sta oy + 1
+    sta count_frames
+    sta count_pixels
+
+    ; total_ms = 0.0; total_pixels = 0.0
+    ldx #total_ms
+    jsr ZF1
+    ldx #total_pixels
+    jsr ZF1
 
     ; zoom = 2x
     lda #1
@@ -675,6 +746,12 @@ copy_byte_loop:
     lda #$22
     sta DMACTL
 
+    ; install the vblank handler
+    lda #7 ; deferred
+    ldx #.hibyte(vblank_handler)
+    ldy #.lobyte(vblank_handler)
+    jsr SETVBV
+
 main_loop:
     ; sy = -92 .. 91
     lda #(256-half_height)
@@ -695,6 +772,108 @@ loop_sx:
     jsr mandelbrot
     jsr pset
 
+
+    ; check if we should update the counters
+    ;
+    ; count_pixels >= width? update!
+    inc count_pixels
+    lda count_pixels
+    cmp #width
+    bmi update_status
+
+    ; count_frames >= 120? update!
+    lda count_frames
+    cmp #120 ; >= 2 seconds
+    bmi skip_status
+
+update_status:
+    ; FR0 = (float)count_pixels & clear count_pixels
+    lda count_pixels
+    sta FR0
+    lda #0
+    sta FR0 + 1
+    sta count_pixels
+    jsr IFP
+
+    ; FR1 = total_pixels
+    ldx #.lobyte(total_pixels)
+    ldy #.hibyte(total_pixels)
+    jsr FLD1R
+
+    ; FR0 += FR1
+    jsr FADD
+
+    ; total_pixels = FR0
+    ldx #.lobyte(total_pixels)
+    ldy #.hibyte(total_pixels)
+    jsr FST0R
+
+
+    ; FR0 = (float)count_frames & clear count_frames
+    ; warning: this should really disable interrupts @TODO
+    lda count_frames
+    sta FR0
+    lda #0
+    sta FR0 + 1
+    sta count_frames
+    jsr IFP
+
+    ; FR0 *= ms_per_frame
+    ldx #.lobyte(ms_per_frame)
+    ldy #.hibyte(ms_per_frame)
+    jsr FLD1R
+    jsr FMUL
+
+    ; FR0 += total_ms
+    ldx #total_ms
+    ldy #0
+    jsr FLD1R
+    jsr FADD
+
+    ; total_ms = FR0
+    ldx #total_ms
+    ldy #0
+    jsr FST0R
+
+    ; FR0 /= total_pixels
+    ldx #total_pixels
+    ldy #0
+    jsr FLD1R
+    jsr FDIV
+
+    ; convert to ASCII in INBUFF
+    jsr FASC
+
+    ; find the last byte
+    ldy #0
+number_loop:
+    lda (INBUFF),y
+    bmi lastchar
+
+    tax
+    lda char_map,x
+    sta textbuffer + speed_start,y
+
+    iny
+    bpl number_loop
+lastchar:
+    ; Y is last char
+    ; trim that high bit
+    and #$7f
+    tax
+    lda char_map,x
+    sta textbuffer + speed_start,y
+
+    ; Fill out any remaining spaces
+    lda #0
+space_loop:
+    iny
+    sta textbuffer + speed_start,y
+    cpy #(20)
+    bmi space_loop
+
+skip_status:
+
     clc
     lda sx
     adc #1