Apple-2-SW/prog8
1
0
Fork 0
mirror of https://github.com/irmen/prog8.git synced 2024-11-26 11:49:22 +00:00
Code Issues Projects Releases Wiki Activity

added some more optimized mul_word asm routines

Browse Source
This commit is contained in:
Irmen de Jong 2020-09-14 22:52:04 +02:00
parent d260182ef3
commit aff6b1fca5
4 changed files with 155 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

142
compiler/res/prog8lib/math.asm
View File

@ -660,26 +660,25 @@ stack_mul_byte_25 .proc
.pend
stack_mul_word_25 .proc
; W + W*8 + W*16
; W = (W*2 + W) *8 + W
lda P8ESTACK_HI+1,x
sta P8ZP_SCRATCH_W1+1
lda P8ESTACK_LO+1,x
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
sta P8ZP_SCRATCH_W1
clc
adc P8ESTACK_LO+1,x
sta P8ESTACK_LO+1,x
sta P8ZP_SCRATCH_W1
lda P8ZP_SCRATCH_W1+1
adc P8ESTACK_HI+1,x
sta P8ESTACK_HI+1,x
sta P8ZP_SCRATCH_W1+1
lda P8ZP_SCRATCH_W1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
clc
adc P8ESTACK_LO+1,x
sta P8ESTACK_LO+1,x
@ -732,7 +731,11 @@ stack_mul_byte_50 .proc
.pend
stack_mul_word_50 .proc
; TODO word * 50
; W = W * 25 * 2
jsr stack_mul_word_25
asl P8ESTACK_LO+1,x
rol P8ESTACK_HI+1,x
rts
.pend
stack_mul_byte_80 .proc
@ -745,7 +748,11 @@ stack_mul_byte_80 .proc
.pend
stack_mul_word_80 .proc
; TODO word * 80
; W = W * 40 * 2
jsr stack_mul_word_40
asl P8ESTACK_LO+1,x
rol P8ESTACK_HI+1,x
rts
.pend
stack_mul_byte_100 .proc
@ -758,7 +765,13 @@ stack_mul_byte_100 .proc
.pend
stack_mul_word_100 .proc
; TODO word * 100
; W = W * 25 * 4
jsr stack_mul_word_25
asl P8ESTACK_LO+1,x
rol P8ESTACK_HI+1,x
asl P8ESTACK_LO+1,x
rol P8ESTACK_HI+1,x
rts
.pend
@ -773,7 +786,8 @@ mul_byte_3 .proc
.pend
mul_word_3 .proc
.warn "todo mul_word_3"
.error "todo mul_word_3"
; AY = AY*2 + AY
rts
.pend
@ -789,7 +803,8 @@ mul_byte_5 .proc
.pend
mul_word_5 .proc
.warn "todo mul_word_5"
.error "todo mul_word_5"
; AY = AY*4 + AY
rts
.pend
@ -805,7 +820,8 @@ mul_byte_6 .proc
.pend
mul_word_6 .proc
.warn "todo mul_word_6"
.error "todo mul_word_6"
; AY = (AY*2 + AY)*2
rts
.pend
@ -821,7 +837,8 @@ mul_byte_7 .proc
.pend
mul_word_7 .proc
.warn "todo mul_word_7"
.error "todo mul_word_7"
; AY = AY*8 - AY
rts
.pend
@ -837,7 +854,8 @@ mul_byte_9 .proc
.pend
mul_word_9 .proc
.warn "todo mul_word_9"
.error "todo mul_word_9"
; AY = AY*8 + AY
rts
.pend
@ -853,7 +871,8 @@ mul_byte_10 .proc
.pend
mul_word_10 .proc
.warn "todo mul_word_10"
.error "todo mul_word_10"
; AY=(AY*4 + AY)*2
rts
.pend
@ -884,7 +903,8 @@ mul_byte_12 .proc
.pend
mul_word_12 .proc
.warn "todo mul_word_12"
.error "todo mul_word_12"
; AY=(AY*2 + AY)*4
rts
.pend
@ -930,7 +950,8 @@ mul_byte_15 .proc
.pend
mul_word_15 .proc
.warn "todo mul_word_15"
.error "todo mul_word_15"
; AY = AY * 16 - AY
rts
.pend
@ -947,7 +968,12 @@ mul_byte_20 .proc
.pend
mul_word_20 .proc
.warn "todo mul_word_20"
; AY = AY * 10 * 2
jsr mul_word_10
sty P8ZP_SCRATCH_REG
asl a
rol P8ZP_SCRATCH_REG
ldy P8ZP_SCRATCH_REG
rts
.pend
@ -966,7 +992,33 @@ mul_byte_25 .proc
.pend
mul_word_25 .proc
.warn "todo mul_word_25"
; AY = (AY*2 + AY) *8 + AY
sta P8ZP_SCRATCH_W1
sty P8ZP_SCRATCH_W1+1
sta P8ZP_SCRATCH_W2
sty P8ZP_SCRATCH_W2+1
asl a
rol P8ZP_SCRATCH_W1+1
clc
adc P8ZP_SCRATCH_W2
sta P8ZP_SCRATCH_W1
lda P8ZP_SCRATCH_W1+1
adc P8ZP_SCRATCH_W2+1
sta P8ZP_SCRATCH_W1+1
lda P8ZP_SCRATCH_W1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
clc
adc P8ZP_SCRATCH_W2
sta P8ZP_SCRATCH_W1
lda P8ZP_SCRATCH_W1+1
adc P8ZP_SCRATCH_W2+1
tay
lda P8ZP_SCRATCH_W1
rts
.pend
@ -979,7 +1031,30 @@ _forties .byte 0*40, 1*40, 2*40, 3*40, 4*40, 5*40, 6*40, 7*40 & 255
.pend
mul_word_40 .proc
.warn "todo mul_word_40"
; AY = (AY*4 + AY)*8
sta P8ZP_SCRATCH_W1
sty P8ZP_SCRATCH_W1+1
sta P8ZP_SCRATCH_W2
sty P8ZP_SCRATCH_W2+1
asl a
rol P8ZP_SCRATCH_W1+1
asl a
rol P8ZP_SCRATCH_W1+1
clc
adc P8ZP_SCRATCH_W2
sta P8ZP_SCRATCH_W1
lda P8ZP_SCRATCH_W1+1
adc P8ZP_SCRATCH_W2+1
asl P8ZP_SCRATCH_W1
rol a
asl P8ZP_SCRATCH_W1
rol a
asl P8ZP_SCRATCH_W1
rol a
asl P8ZP_SCRATCH_W1
rol a
tay
lda P8ZP_SCRATCH_W1
rts
.pend
@ -992,7 +1067,12 @@ _fifties .byte 0*50, 1*50, 2*50, 3*50, 4*50, 5*50, 6*50 & 255, 7*50 & 255
.pend
mul_word_50 .proc
.warn "todo mul_word_50"
; AY = AY * 25 * 2
jsr mul_word_25
sty P8ZP_SCRATCH_REG
asl a
rol P8ZP_SCRATCH_REG
ldy P8ZP_SCRATCH_REG
rts
.pend
@ -1005,7 +1085,12 @@ _eighties .byte 0*80, 1*80, 2*80, 3*80
.pend
mul_word_80 .proc
.warn "todo mul_word_80"
; AY = AY * 40 * 2
jsr mul_word_40
sty P8ZP_SCRATCH_REG
asl a
rol P8ZP_SCRATCH_REG
ldy P8ZP_SCRATCH_REG
rts
.pend
@ -1018,7 +1103,14 @@ _hundreds .byte 0*100, 1*100, 2*100, 3*100 & 255
.pend
mul_word_100 .proc
.warn "todo mul_word_100"
; AY = AY * 25 * 4
jsr mul_word_25
sty P8ZP_SCRATCH_REG
asl a
rol P8ZP_SCRATCH_REG
asl a
rol P8ZP_SCRATCH_REG
ldy P8ZP_SCRATCH_REG
rts
.pend

4
compiler/src/prog8/compiler/target/c64/codegen/assignment/AugmentableAssignmentAsmGen.kt
View File

@ -692,13 +692,13 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
"*" -> {
if(dt == DataType.UWORD){
if(value in asmgen.optimizedWordMultiplications) {
TODO("optimized var uword mul litval $value")
asmgen.out(" lda $name | ldy $name+1 | jsr math.mul_word_$value | sta $name | sty $name+1")
} else {
TODO("var uword mul litval $value")
}
} else {
if(value.absoluteValue in asmgen.optimizedWordMultiplications) {
TODO("optimized var sword mul litval $value")
asmgen.out(" lda $name | ldy $name+1 | jsr math.mul_word_$value | sta $name | sty $name+1")
} else {
// TODO don't use stack here
// TODO does this work for signed words?

1
docs/source/todo.rst
View File

@ -2,6 +2,7 @@
TODO
====
- compiler errors and warnings in standard format so the IDE shows them as clickable links; ./test.asm:2578:3: blablabla
- further optimize assignment codegeneration
- auto select correct library to import based on target, instead of having c64- and cx16- prefix variants
- get rid of all TODO's ;-)

49
examples/test.p8
View File

@ -10,23 +10,44 @@ main {
sub start() {
ubyte ub1
ubyte ii
for ii in 0 to 10 {
uword ub1
word ww1
uword ii
for ii in 0 to 20 {
; ub1 = ii
; ub1 *= 40 ; TODO implement non-stack optimized muls
; todo a = EXPRESSION * const -> is that optimized?
ub1 = 1+ii * 40
txt.print_ub(ub1)
c64.CHROUT(',')
ub1 = 1+ii * 50
txt.print_ub(ub1)
c64.CHROUT(',')
ub1 = 1+ii * 80
txt.print_ub(ub1)
c64.CHROUT(',')
ub1 = 1+ii * 100
txt.print_ub(ub1)
ub1 = ii * 25
txt.print_uw(ub1)
; c64.CHROUT(',')
; ub1 = ii * 50
; txt.print_uw(ub1)
; c64.CHROUT(',')
; ub1 = ii * 80
; txt.print_uw(ub1)
; c64.CHROUT(',')
; ub1 = ii * 100
; txt.print_uw(ub1)
c64.CHROUT('\n')
}
c64.CHROUT('\n')
for ii in 0 to 20 {
; ub1 = ii
; ub1 *= 40 ; TODO implement non-stack optimized muls
; todo a = EXPRESSION * const -> is that optimized?
ww1 = (-ii) * 25
txt.print_w(ww1)
; c64.CHROUT(',')
; ub1 = ii * 50
; txt.print_uw(ub1)
; c64.CHROUT(',')
; ub1 = ii * 80
; txt.print_uw(ub1)
; c64.CHROUT(',')
; ub1 = ii * 100
; txt.print_uw(ub1)
c64.CHROUT('\n')
}