MMX: fast integer alpha blending

blend_mmx.asm

macro CANVAS_SET_ALPHA col {	; prepare mmx registers (alpha) by col (0xAAxxxxxx)
	push	col
	shr	col, 24
	pxor	mm5, mm5
	movd	mm7, col
	pshufw	mm7, mm7, 0
	neg	col
	add	col, 256
	movd	mm6, col
	pshufw	mm6, mm6, 0
	pop	col
}

; Result := (Frgb*Fa + Brgb*(256-Fa))/256
macro CANVAS_SET_PIXEL dst_ptr, color {
	movd		mm0, [dst_ptr]	; mm0(A) = 0 0 0 0 | 0 Ra Ga Ba
	movd		mm1, color	; mm1(B) = 0 0 0 0 | 0 Rb Gb Bb
	punpcklbw	mm0, mm5	; mm0 = 0 0 0 Ra | 0 Ga 0 Ba
	punpcklbw	mm1, mm5	; mm1 = 0 0 0 Rb | 0 Gb 0 Bb
	pmullw		mm0, mm6	; mm0 = 0 Ra*X | Ga*X Ba*X
	pmullw		mm1, mm7	; mm1 = 0 Rb*Y | Gb*Y Bb*Y
	paddusw		mm0, mm1	; mm0 = 0 Ra*X+Rb*y | Ga*X+Gb*y Ba*X+Bb*Y
	psrlw		mm0, 8		; mm0 = 0 0 0 Rc | 0 Gc 0 Bc
	packuswb	mm0, mm0	; mm0 = 0 0 0 0 | 0 Rc Gc Bc
	movd		[dst_ptr], mm0
}
...
mov eax, 0x800000ff		; red color with 50% alpha
CANVAS_SET_ALPHA eax		; prepare MMX alpha-registers
CANVAS_SET_PIXELA esi,eax	; blend pixel from [esi] to eax by alpha, and write it to [esi]