izaya/collapseos
1
0
Fork 0
mirror of https://github.com/hsoft/collapseos.git synced 2025-04-05 06:48:39 +11:00
Code Issues Releases Wiki Activity

Removed skip leading zeroes, added skip first multiply

Browse Source 
Now instead of skipping leading zeroes, the first digit is loaded directly into hl without first multiplying by 10. This means the first loop is skipped in the overhead, making the method 2-3 times faster overall, and is now faster for the more common fewer digit cases too. The number of bytes is exactly the same, and the inner loop is slightly faster too thanks to no longer needing to load a into c.
To be more precise about the speed increase over the current code, for decimals of length 1 it'll be 3.18x faster, for decimals of length 2, 2.50x faster, for length 3, 2.31x faster, for length 4, 2.22x faster, and for length 5 and above, at least 2.03x faster. In terms of cycles, this is around 100+(132*length) cycles saved per decimal.
This commit is contained in:
Clanmaster21 2019-10-20 22:26:07 +01:00 committed by GitHub
parent ac2d899e80
commit 67adc6fcfc
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 21 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
apps/lib/parse.asm
View File

@ -17,29 +17,32 @@
; same conditions as parseLiteral.
; Sets Z on success, unset on error.
; 55 bytes, 32 cycles in first loop
; 90 cycles overhead + up to 69 cycles if length >= 5
; 140 cycles in loop
; 55 bytes
; 107 cycles overhead + up to 45 cycles if length >= 5
; 136 cycles in loop
; first digit is skipped in overhead
parseDecimal:
push hl
.skip: ; Skips leading zeroes
ld a, (hl)
inc hl
cp '0'
jr z, .skip
ld a, (hl)
add a, 0xc6 ; converts '0'-'9' to 0-9
sub 0xf6 ; carries if out of range
exx ; preserve bc, hl, de
ld hl, 0
ld b, 5 ; Carries can only occur for decimals >=5 in length
jr .start
ld h, 0
ld l, a ; load first digit in without multiplying
ld b, 4 ; Carries can only occur for decimals >=5 in length
jr c, .error
.loop:
ld c, a ; c holds current digit
exx ;swap hl back in to get address
ld a, (hl) ; a checks if following digit is null at end of loop
inc hl
exx
inc hl
ld a, (hl)
exx
add a, 0xc6 ; converts '0'-'9' to 0-9
sub 0xf6 ; carries if out of range
jr c, .error
add hl, hl ; x2
ld d, h
ld e, l ; de is x2
@ -47,16 +50,9 @@ parseDecimal:
add hl, hl ; x8
add hl, de ; x10
ld d, 0
ld e, c
ld e, a
add hl, de
jr c, .error ; if hl was 0x1999, it may carry here
; This check could be taken outside the loop, but at the cost of 6 bytes
.start:
add a, 0xc6 ; converts '0'-'9' to 0-9
sub 0xf6 ; carries if out of range
jr c, .error
djnz .loop
@ -71,8 +67,7 @@ parseDecimal:
sub 0xd0 ; if a is null, set Z
; a is checked for null before any errors
.end:
push hl
pop ix
push hl \ pop ix
exx ; restore original de and bc
pop hl
ret