... : PAPER 0: INK 7: PAUSE 1: BORDER 0: CLS
... p$(50,1): FOR a=1 TO 50: LET p$(a)=CHR$ (65 ... $=p$(a): LET b=a-1: LET done=0
... p$(b)=p$(b+1): LET b=b+1: GO TO 210 ... $(b)=t$: LET a=a-1: GO TO 190
... .6% better than Bubble Sort 1.

Can anyone ...

... : PAPER 0: INK 7: PAUSE 1: BORDER 0: CLS
... p$(50,1): FOR a=1 TO 50: LET p$(a)=CHR$ (65 ... $=p$(a): LET b=a-1: LET done=0
... p$(b)=p$(b+1): LET b=b+1: GO TO 210 ... $(b)=t$: LET a=a-1: GO TO 190
... .6% better than Bubble Sort 1.

Can anyone ...

... : PAPER 0: INK 7: PAUSE 1: BORDER 0: CLS
110 ... p$(50,1): FOR a=1 TO 50: LET p$(a)=CHR$ (65 ...

Hi Dr. Beep,

Tried one of the tricks you said (move the code nearer the top) and a couple of other optimisations and the Shell Sort now averages 10.3 seconds rather than 11.57.

[CODE]100 GO SUB 250
110 REM SHELL SORT V2

...
[code]
1 REM Shell Sort
10 ... FOR i=1 TO n:
LET v$(i)=CHR$ (65 ... +RND*25):
NEXT i
20 FOR i=1 ... br /> LET m=(m-1)/2:
GO ... -st:
FOR i=1 TO n:
PRINT ...

... FOR i=1 TO 52:
LET v$(i)=CHR$ (i+64 ... ADDRESS
ld a,(IX+1) ; =l$
ld (hl ... nz,FORS
ld a,1
cp (IX+2 ... ret z ;ret if m=1
dec (IX+2)

Hi Battle Bunny,

Never ceases to amaze me the difference in performance between BASIC and MCODE. Sounds like you have a powerful routine there.

I have modified my demo version so that it can be used (for future generations ;)) as ...

... br />
[CODE]
1 LET p=3:
POKE ... 1 TO 7:
PRINT AT y*3-2,23;CHR ... y*3;CHR$ (71+y);AT 21,y*3+1;CHR$ (78 ... +y); PAPER 1; AT x*3 ... 5-p:
BORDER p+1:
INPUT ':
... +NOT t(x):
BEEP 1,9
[/CODE]

... quicksort implementation...
[code] 1 DEF FN t()=(PEEK 23672+ ... $(nelem,1): FOR n=1 TO nelem: LET
p$(n)=CHR$ ( ... STOP
1000 FOR n=1 TO nelem: PRINT p$(n ... (sp,3): LET sp=sp-1: RETURN
4000 LET ... /> 4030 IF end-pv>1 THEN GO SUB 2000: LET ...