mandel-6502/mandel.s

384 lines
7.9 KiB
ArmAsm
Raw Normal View History

; Our zero-page vars
sx = $80 ; 8 bits: screen pixel x
sy = $81 ; 8 bits: screen pixel y
cx = $82 ; 16 bits fixed point
cy = $84 ; 16 bits fixed point
zx = $86 ; 16 bits fixed point
zy = $88 ; 16 bits fixed point
zx_2 = $8a ; 32 bits fixed point
zy_2 = $8e ; 32 bits fixed point
zx_zy = $92 ; 32 bits fixed point
dist = $96 ; 32 bits fixed point
iter = $9a ; 8 bits iteration count
temp = $a0 ; debug temp area
2022-12-29 05:08:16 +00:00
; FP registers in zero page
FR0 = $d4
FRE = $da
FR1 = $e0
FR2 = $e6
.code
.export start
; 2 + 9 * byte cycles
.macro add bytes, dest, arg1, arg2
clc ; 2 cyc
.repeat bytes, byte ; 9 * byte cycles
lda arg1 + byte
adc arg2 + byte
sta dest + byte
.endrepeat
.endmacro
.macro add16 dest, arg1, arg2
add 2, dest, arg1, arg2
.endmacro
.macro add32 dest, arg1, arg2
add 2, dest, arg2, dest
.endmacro
; 2 + 9 * byte cycles
.macro sub bytes, dest, arg1, arg2
sec ; 2 cyc
.repeat bytes, byte ; 9 * byte cycles
lda arg1 + byte
sbc arg2 + byte
sta dest + byte
.endrepeat
.endmacro
.macro sub16 dest, arg1, arg2
sub 2, dest, arg1, arg2
.endmacro
.macro sub32 dest, arg1, arg2
sub 4, dest, arg1, arg2
.endmacro
.macro shl bytes, arg
asl arg
.repeat bytes-1
rol arg
.endrepeat
.endmacro
.macro shl16 arg
shl 2, arg
.endmacro
.macro shl24 arg
shl 3, arg
.endmacro
.macro shl32 arg
shl 4, arg
.endmacro
; 6 * bytes cycles
.macro copy bytes, dest, arg
.repeat bytes, byte ; 6 * bytes cycles
lda arg + byte ; 3 cyc
sta dest + byte ; 3 cyc
.endrepeat
.endmacro
.macro copy16 dest, arg
copy 2, dest, arg
.endmacro
.macro copy32 dest, arg
copy 4, dest, arg
.endmacro
2022-12-30 08:43:44 +00:00
; 2 + 8 * byte cycles
.macro neg bytes, arg
sec ; 2 cyc
.repeat bytes, byte ; 8 * byte cycles
lda #00 ; 2 cyc
sbc arg + byte ; 3 cyc
sta arg + byte ; 3 cyc
.endrepeat
.endmacro
; 18 cycles
.macro neg16 arg
neg 2, arg
.endmacro
; 34 cycles
.macro neg32 arg
neg 4, arg
.endmacro
2022-12-31 02:25:43 +00:00
; inner loop for imul16
2023-01-05 04:21:51 +00:00
; bitnum < 8: 25 or 41 cycles
; bitnum >= 8: 30 or 46 cycles
2022-12-30 04:18:21 +00:00
.macro bitmul16 arg1, arg2, result, bitnum
2023-01-05 03:52:56 +00:00
.local zero
2023-01-05 04:12:34 +00:00
.local one
.local next
2022-12-29 11:37:51 +00:00
2022-12-30 08:43:44 +00:00
; does 16-bit adds
2023-01-05 05:09:45 +00:00
; arg1 and arg2 are treated as unsigned
; negative signed inputs must be flipped first
2022-12-30 08:43:44 +00:00
2023-01-05 04:12:34 +00:00
; 7 cycles up to the branch
2022-12-30 04:18:21 +00:00
; check if arg1 has 0 or 1 bit in this place
2022-12-31 01:33:18 +00:00
; 5 cycles either way
2022-12-30 04:18:21 +00:00
.if bitnum < 8
2022-12-31 01:33:18 +00:00
lda arg1 ; 3 cyc
and #(1 << bitnum) ; 2 cyc
2022-12-30 04:18:21 +00:00
.else
2022-12-31 01:33:18 +00:00
lda arg1 + 1 ; 3 cyc
and #(1 << (bitnum - 8)) ; 2 cyc
2022-12-30 04:18:21 +00:00
.endif
2023-01-05 04:12:34 +00:00
bne one ; 2 cyc
2023-01-05 04:21:51 +00:00
zero: ; 18 cyc, 23 cyc
2023-01-05 04:12:34 +00:00
lsr result + 3 ; 5 cyc
jmp next ; 3 cyc
2022-12-29 11:37:51 +00:00
2023-01-05 04:21:51 +00:00
one: ; 32 cyc, 37 cyc
2022-12-29 11:37:51 +00:00
; 16-bit add on the top bits
2023-01-05 04:12:34 +00:00
clc ; 2 cyc
2022-12-31 02:21:31 +00:00
lda result + 2 ; 3 cyc
adc arg2 ; 3 cyc
sta result + 2 ; 3 cyc
lda result + 3 ; 3 cyc
adc arg2 + 1 ; 3 cyc
2023-01-05 05:09:45 +00:00
ror a ; 2 cyc - get a jump on the shift
2022-12-31 02:21:31 +00:00
sta result + 3 ; 3 cyc
next:
2022-12-31 01:33:18 +00:00
ror result + 2 ; 5 cyc
ror result + 1 ; 5 cyc
2022-12-30 04:18:21 +00:00
.if bitnum >= 8
2023-01-05 04:21:51 +00:00
; we can save 5 cycles * 8 bits = 40 cycles total by skipping this byte
; when it's all uninitialized data
2022-12-31 01:33:18 +00:00
ror result ; 5 cyc
2022-12-30 04:18:21 +00:00
.endif
2023-01-05 04:21:51 +00:00
2023-01-05 04:12:34 +00:00
2022-12-29 05:08:16 +00:00
.endmacro
2022-12-31 01:33:18 +00:00
; 5 to 25 cycles
2022-12-30 08:43:44 +00:00
.macro check_sign arg
; Check sign bit and flip argument to postive,
; keeping a count of sign bits in the X register.
.local positive
2022-12-31 01:33:18 +00:00
lda arg + 1 ; 3 cyc
bpl positive ; 2 cyc
neg16 arg ; 18 cyc
inx ; 2 cyc
2022-12-30 08:43:44 +00:00
positive:
.endmacro
2022-12-29 05:08:16 +00:00
; 518 - 828 cyc
.macro imul16 dest, arg1, arg2
copy16 FR0, arg1 ; 12 cyc
copy16 FR1, arg2 ; 12 cyc
jsr imul16_func ; 470-780
copy32 dest, FR2 ; 24 cyc
.endmacro
2023-01-05 04:37:16 +00:00
; min 470 cycles
; max 780 cycles
.proc imul16_func
2022-12-30 08:43:44 +00:00
arg1 = FR0 ; 16-bit arg (clobbered)
arg2 = FR1 ; 16-bit arg (clobbered)
result = FR2 ; 32-bit result
2022-12-31 01:33:18 +00:00
ldx #0 ; 2 cyc
2022-12-30 08:43:44 +00:00
; counts the number of sign bits in X
2022-12-31 01:33:18 +00:00
check_sign arg1 ; 5 to 25 cyc
check_sign arg2 ; 5 to 25 cyc
2022-12-30 08:43:44 +00:00
2022-12-30 04:18:21 +00:00
; zero out the 32-bit temp's top 16 bits
2022-12-31 01:33:18 +00:00
lda #0 ; 2 cyc
sta result + 2 ; 3 cyc
sta result + 3 ; 3 cyc
2022-12-29 11:37:51 +00:00
; the bottom two bytes will get cleared by the shifts
2022-12-29 05:08:16 +00:00
2022-12-30 08:43:44 +00:00
; unrolled loop for maximum speed, at the cost
; of a larger routine
2023-01-05 04:37:16 +00:00
; 440 to 696 cycles
2022-12-29 05:08:16 +00:00
.repeat 16, bitnum
2023-01-05 04:37:16 +00:00
; bitnum < 8: 25 or 41 cycles
; bitnum >= 8: 30 or 46 cycles
2022-12-30 08:43:44 +00:00
bitmul16 arg1, arg2, result, bitnum
2022-12-29 05:08:16 +00:00
.endrepeat
2022-12-30 04:18:21 +00:00
2022-12-30 08:43:44 +00:00
; In case of mixed input signs, return a negative result.
2022-12-31 01:33:18 +00:00
cpx #1 ; 2 cyc
bne positive_result ; 2 cyc
neg32 result ; 34 cyc
2022-12-30 08:43:44 +00:00
positive_result:
2022-12-31 01:33:18 +00:00
rts ; 6 cyc
2022-12-29 05:08:16 +00:00
.endproc
.macro round16 arg
; Round top 16 bits of 32-bit fixed-point number in-place
.local zero
.local one
.local positive
.local negative
.local neg2
.local next
; no round - 5 cycles
; round pos, no carry - 17
; round pos, carry - 22
; round neg, no carry - 23
; round neg, carry - 28
; average = 5 / 2 + (17 + 22 + 23 + 28) / 8
; = 5 / 2 + 90 / 8
; = 2.5 + 11.25 = 13.75 cycles average on evenly distributed input
lda arg + 1 ; 3 cyc
bpl zero ; 2 cyc
one:
; check sign bit
lda arg + 3 ; 3 cyc
bpl positive ; 2 cyc
negative:
lda arg + 2 ; 3 cyc
beq neg2 ; 2 cyc
dec arg + 2 ; 5 cyc
jmp next ; 3 cyc
neg2:
dec arg + 2 ; 5 cyc
dec arg + 3 ; 5 cyc
jmp next ; 3 cyc
positive:
inc arg + 2 ; 5 cyc
beq next ; 2 cyc
inc arg + 3 ; 5 cyc
zero:
next:
.endmacro
.proc mandelbrot
; input:
; cx: position scaled to 4.12 fixed point - -8..+7.9
; cy: position scaled to 4.12
;
; output:
; iter: iteration count at escape or 0
2022-12-29 05:08:16 +00:00
; zx = 0
; zy = 0
; zx_2 = 0
2022-12-30 08:55:48 +00:00
; zy_2 = 0
; zx_zy = 0
; dist = 0
; iter = 0
lda #00
ldx iter - zx
initloop:
sta zx,x
dex
bne initloop
2022-12-29 05:08:16 +00:00
loop:
; 1939 - 3007 cyc
; iter++ & max-iters break = 7 cyc
inc iter ; 5 cyc
bne keep_going ; 2 cyc
rts
keep_going:
2022-12-29 05:08:16 +00:00
; 4.12: (-8 .. +7.9)
; zx = zx_2 - zy_2 + cx = 3 * 20 = 60 cyc
sub16 zx, zx_2, zy_2
add16 zx, zx, cx
; zy = zx_zy + zx_zy + cy = 3 * 20 = 60 cyc
sub16 zy, zx_zy, zx_zy
add16 zy, zy, cy
2022-12-29 05:08:16 +00:00
2023-01-05 19:58:32 +00:00
; 8.24: (-128 .. +127.9)
; zx_2 = zx * zx = 518 - 828 cyc
imul16 zx_2, zx, zx
; zy_2 = zy * zy = 518 - 828 cyc
imul16 zy_2, zy, zy
; zx_zy = zx * zy = 518 - 828 cyc
imul16 zx_zy, zx, zy
; dist = zx_2 + zy_2 = 38 cyc
add32 dist, zx_2, zy_2
; if dist >= 4 break, else continue iterating = 7 cyc
lda dist + 3 ; 3 cyc
cmp #4 ; 2 cyc
bmi still_in ; 2 cyc
rts
still_in:
; shift and round zx_2 to 4.12 = (60 + 5) - (60 + 28) = 65 - 88 cyc
.repeat 4 ; 60 cyc
shl24 zx_2 ; 15 cyc
.endrepeat
round16 zx_2 ; 5-28 cycles
2022-12-29 05:08:16 +00:00
; shift and round zy_2 to 4.12 = (20 + 5) - (20 + 28) = 65 - 88 cyc
.repeat 4 ; 60 cyc
shl24 zy_2 ; 15 cyc
.endrepeat
round16 zy_2 ; 5-28 cycles
; shift and round zx_zy to 4.12 = (20 + 5) - (20 + 28) = 65 - 88 cyc
.repeat 4 ; 60 cyc
shl24 zx_zy ; 15 cyc
.endrepeat
round16 zx_zy ; 5-28 cycles
2022-12-30 08:55:48 +00:00
; if may be in the lake, look for looping output with a small buffer
; as an optimization vs running to max iters
jmp loop ; 3 cycles
2022-12-30 08:55:48 +00:00
2022-12-29 05:08:16 +00:00
.endproc
2022-12-30 04:32:58 +00:00
.proc start
2022-12-30 08:43:44 +00:00
2023-01-05 04:12:34 +00:00
looplong:
; cx = -0.5
lda #$f7
sta cx
2022-12-30 08:43:44 +00:00
lda #$ff
sta cx + 1
2022-12-30 04:32:58 +00:00
; cy = 1
lda #$10
sta cy
lda #$00
sta cy + 1
jsr mandelbrot
2022-12-30 08:43:44 +00:00
; should have 32-bit -15 in FR2
2022-12-30 04:32:58 +00:00
; save the completed iter count for debugging
lda iter
sta temp
2023-01-05 04:12:34 +00:00
loop:
; keep looping over so we can work in the debugger
jmp looplong
2022-12-30 04:32:58 +00:00
.endproc