mirror of
https://github.com/irmen/prog8.git
synced 2024-11-27 03:50:27 +00:00
274 lines
5.4 KiB
Lua
274 lines
5.4 KiB
Lua
floats {
|
|
; the floating point functions shared across compiler targets
|
|
%option merge, no_symbol_prefixing, ignore_unused
|
|
|
|
const float π = 3.141592653589793
|
|
const float PI = π
|
|
const float TWOPI = 2*π
|
|
|
|
|
|
asmsub print(float value @FAC1) clobbers(A,X,Y) {
|
|
; ---- prints the floating point value (without a newline). No leading space (unlike BASIC)!
|
|
%asm {{
|
|
jsr tostr
|
|
ldy #0
|
|
- lda (P8ZP_SCRATCH_W1),y
|
|
beq +
|
|
jsr cbm.CHROUT
|
|
iny
|
|
bne -
|
|
+ rts
|
|
}}
|
|
}
|
|
|
|
asmsub tostr(float value @FAC1) clobbers(X) -> str @AY {
|
|
; ---- converts the floating point value to a string. No leading space!
|
|
%asm {{
|
|
jsr FOUT
|
|
sta P8ZP_SCRATCH_W1
|
|
sty P8ZP_SCRATCH_W1+1
|
|
ldy #0
|
|
lda (P8ZP_SCRATCH_W1),y
|
|
cmp #' '
|
|
bne +
|
|
inc P8ZP_SCRATCH_W1
|
|
bne +
|
|
inc P8ZP_SCRATCH_W1+1
|
|
+ lda P8ZP_SCRATCH_W1
|
|
ldy P8ZP_SCRATCH_W1+1
|
|
rts
|
|
}}
|
|
}
|
|
|
|
sub pow(float value, float power) -> float {
|
|
%asm {{
|
|
stx P8ZP_SCRATCH_W1
|
|
sty P8ZP_SCRATCH_W1+1
|
|
lda #<value
|
|
ldy #>value
|
|
jsr floats.CONUPK
|
|
lda #<power
|
|
ldy #>power
|
|
jsr floats.FPWR
|
|
ldx P8ZP_SCRATCH_W1
|
|
ldy P8ZP_SCRATCH_W1+1
|
|
rts
|
|
}}
|
|
}
|
|
|
|
sub sin(float angle) -> float {
|
|
%asm {{
|
|
lda #<angle
|
|
ldy #>angle
|
|
jsr MOVFM
|
|
jmp SIN
|
|
}}
|
|
}
|
|
|
|
sub cos(float angle) -> float {
|
|
%asm {{
|
|
lda #<angle
|
|
ldy #>angle
|
|
jsr MOVFM
|
|
jmp COS
|
|
rts
|
|
}}
|
|
}
|
|
|
|
sub tan(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jmp TAN
|
|
}}
|
|
}
|
|
|
|
sub atan(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jmp ATN
|
|
}}
|
|
}
|
|
|
|
; two-argument arctangent that returns an angle in the correct quadrant
|
|
; for the signs of x and y, normalized to the range [0, 2π]
|
|
sub atan2(float y, float x) -> float {
|
|
float atn
|
|
if x == 0 {
|
|
atn = π/2
|
|
if y == 0 return 0
|
|
if y < 0 {
|
|
atn += π
|
|
}
|
|
} else {
|
|
atn = atan(y / x)
|
|
}
|
|
if x < 0 atn += π
|
|
if atn < 0 atn += 2*π
|
|
return atn
|
|
}
|
|
|
|
; reciprocal functions
|
|
sub secant(float value) -> float { return 1.0 / cos(value) }
|
|
sub csc(float value) -> float { return 1.0 / sin(value) }
|
|
sub cot(float value) -> float { return 1.0 / tan(value) }
|
|
|
|
sub ln(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jmp LOG
|
|
}}
|
|
}
|
|
|
|
sub log2(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jsr LOG
|
|
jsr MOVEF
|
|
lda #<FL_LOG2_const
|
|
ldy #>FL_LOG2_const
|
|
jsr MOVFM
|
|
jmp FDIVT
|
|
}}
|
|
}
|
|
|
|
sub rad(float angle) -> float {
|
|
; -- convert degrees to radians (d * pi / 180)
|
|
%asm {{
|
|
lda #<angle
|
|
ldy #>angle
|
|
jsr MOVFM
|
|
lda #<_pi_div_180
|
|
ldy #>_pi_div_180
|
|
jmp FMULT
|
|
_pi_div_180 .byte 123, 14, 250, 53, 18 ; pi / 180
|
|
}}
|
|
}
|
|
|
|
sub deg(float angle) -> float {
|
|
; -- convert radians to degrees (d * (1/ pi * 180))
|
|
%asm {{
|
|
lda #<angle
|
|
ldy #>angle
|
|
jsr MOVFM
|
|
lda #<_one_over_pi_div_180
|
|
ldy #>_one_over_pi_div_180
|
|
jmp FMULT
|
|
rts
|
|
_one_over_pi_div_180 .byte 134, 101, 46, 224, 211 ; 1 / (pi * 180)
|
|
}}
|
|
}
|
|
|
|
sub round(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jsr FADDH
|
|
jmp INT
|
|
}}
|
|
}
|
|
|
|
sub floor(float value) -> float {
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
jmp INT
|
|
}}
|
|
}
|
|
|
|
sub ceil(float value) -> float {
|
|
; -- ceil: tr = int(f); if tr==f -> return else return tr+1
|
|
%asm {{
|
|
lda #<value
|
|
ldy #>value
|
|
jsr MOVFM
|
|
ldx #<fmath_float1
|
|
ldy #>fmath_float1
|
|
jsr MOVMF
|
|
jsr INT
|
|
lda #<fmath_float1
|
|
ldy #>fmath_float1
|
|
jsr FCOMP
|
|
cmp #0
|
|
beq +
|
|
lda #<FL_ONE_const
|
|
ldy #>FL_ONE_const
|
|
jsr FADD
|
|
+ rts
|
|
}}
|
|
}
|
|
|
|
sub rndseed(float seed) {
|
|
if seed>0
|
|
seed = -seed ; make sure fp seed is always negative
|
|
|
|
%asm {{
|
|
lda #<seed
|
|
ldy #>seed
|
|
jsr MOVFM ; load float into fac1
|
|
lda #-1
|
|
jmp floats.RND
|
|
}}
|
|
}
|
|
|
|
|
|
sub minf(float f1, float f2) -> float {
|
|
if f1<f2
|
|
return f1
|
|
return f2
|
|
}
|
|
|
|
|
|
sub maxf(float f1, float f2) -> float {
|
|
if f1>f2
|
|
return f1
|
|
return f2
|
|
}
|
|
|
|
|
|
sub clampf(float value, float minimum, float maximum) -> float {
|
|
if value>maximum
|
|
value=maximum
|
|
if value>minimum
|
|
return value
|
|
return minimum
|
|
}
|
|
|
|
inline asmsub push(float value @FAC1) {
|
|
%asm {{
|
|
jsr floats.pushFAC1
|
|
}}
|
|
}
|
|
|
|
inline asmsub pop() -> float @FAC1 {
|
|
%asm {{
|
|
clc
|
|
jsr floats.popFAC
|
|
}}
|
|
}
|
|
|
|
sub lerp(float v0, float v1, float t) -> float {
|
|
; Linear interpolation (LERP)
|
|
; Precise method, which guarantees v = v1 when t = 1.
|
|
; returns an interpolation between two inputs (v0, v1) for a parameter t in the closed unit interval [0, 1]
|
|
return (1 - t) * v0 + t * v1
|
|
}
|
|
|
|
sub lerp_fast(float v0, float v1, float t) -> float {
|
|
; Linear interpolation (LERP)
|
|
; Imprecise (but slightly faster) method, which does not guarantee v = v1 when t = 1
|
|
; returns an interpolation between two inputs (v0, v1) for a parameter t in the closed unit interval [0, 1]
|
|
return v0 + t * (v1 - v0)
|
|
}
|
|
|
|
}
|