Assember: 16bit bubble sort

Here is also 8bit version:

        ;Input BC = Number of bytes to sort (Min 2)

        LD BC,DATAE-DATAS

BUBBLE8:
        DEC BC
        LD IYH,B
        LD IYL,C
.LOOPO
        LD IXL,1
        LD B,IYH
        LD C,IYL
        LD DE,DATAS    ; Data to be sorted out
        LD H,D
        LD L,E
        INC HL
.LOOPI
        LD A,(DE)
        CP (HL)
        JR C,.SWAP
        INC DE
        CPI
        JP PE,.LOOPI
        JP .CHECK
.SWAP
        LDI
        LD (DE),A
        LD IXL,0
        JP NZ,.LOOPI
.CHECK:
        LD A,IXL
        AND A
        RET NZ
        JP .LOOPO


; Example data:

DATAS:   DB 0,1,2,3,4,5,4,3,2,1,0
DATAE:

Bubble sort is the slowest possible sorting algorithm that doesn't perform any steps/swaps that would decrease the "sortedness" of the array of elements. I remember this being proven on some basic computer science lecture...

Bogo sort is even slower, but it's based on randomly shuffling the elements until they're in the wanted order, so it performs those harmful swaps that bubble sort doesn't.
http://en.wikipedia.org/wiki/Bogosort

NYYRIKKI's sorting code looks quite small, though.

Your absolutely right, it all comes down to the amount of data you have to sort !
If you keep track of the highest inserted value (at the end of the new table), yo can make a simple decision if it is needed to search for a specific value or just add the new value at the end of the table

Hmm... I just noticed that there is bug in the 8bit-version. After LDI there should be ofcource JP PE and not JP NZ.

Here is "Insertion sort" version of the 8bit version:

	;Input BC = Number of bytes to sort (Min 2)

	LD BC,DATAE-DATAS

INSERT8:
	EXX
	LD BC,0
.LOOPO
	INC BC
	PUSH BC
	EXX
	CPI
	JP PO,.EXIT
	LD HL,DATAS	; Data to be sorted out
	ADD HL,BC
	PUSH HL
	EXX
	POP HL
	LD D,H
	LD E,L
	DEC DE

.LOOPI
	LD A,(DE)
	CP (HL)
	JR C,.SWAP
	INC DE
	CPI
	JP PE,.LOOPI
	POP BC
	JP .LOOPO

.SWAP
	LDI
	LD (DE),A
	JP PE,.LOOPI
	POP BC
	JP .LOOPO

.EXIT	RET


; Example data:

DATAS:	DB 0,1,2,3,4,5,4,3,2,1,9
DATAE:

If you want it to be fast it's better to forget about the bobble sort in total, try using the insertion sort.
The insertion sort is also a simple (also not the quickest) method that is a lot faster then the bubble sort.
Depends. For very small arrays, bubble sort is the fastest. That's why it is or could be used for supersort.

Under certain conditions, counting-sort is the quickest, time complexity O(n) (insertion-sort and bubble-sort are O(n^2) ).
Insertion and bubble are in-place sorting algorithms, while counting requires additional memory space, but it is stable so in theory it provides benefits to sprite sorting (same shape but different sprite identity).

Never heard of counting sort, at least not under this name. What do you mean by ''in place''? There's something called topological sorting, which is ''in place''. Bubble sort exchanges elements, so they don't stay in place. For sorting an array of n elements, you at least need O(n log n) steps.