sample code of generation ARM NEON instruction

neon_test.c

/*
arm-linux-gnueabi-gcc-4.6 -O2 -march=armv7-a -mtune=cortex-a9 -ftree-vectorize -mhard-float -mfloat-abi=softfp -mfpu=neon -ffast-math -mvectorize-with-neon-quad -S neon_test.c
*/
void NeonTest(short int * __restrict a, short int * __restrict b, short int * __restrict z)
{
  int i;
  for (i = 0; i < 200; i++) {
    z[i] = a[i] * b[i];
  }
}

neon_test.s

	.syntax unified
	.arch armv7-a
	.eabi_attribute 27, 3
	.fpu neon
	.eabi_attribute 23, 1
	.eabi_attribute 24, 1
	.eabi_attribute 25, 1
	.eabi_attribute 26, 2
	.eabi_attribute 30, 2
	.eabi_attribute 34, 1
	.eabi_attribute 18, 4
	.thumb
	.file	"neon_test2.c"
	.text
	.align	2
	.global	NeonTest
	.thumb
	.thumb_func
	.type	NeonTest, %function
NeonTest:
	@ args = 0, pretend = 0, frame = 8
	@ frame_needed = 0, uses_anonymous_args = 0
	@ link register save eliminated.
	ubfx	r3, r0, #1, #3
	push	{r4, r5, r6, r7, r8, r9, sl, fp}
	negs	r3, r3
	sub	sp, sp, #8
	ands	r3, r3, #7
	beq	.L8
	movs	r6, #0
	mov	r4, r6
.L3:
	ldrh	r7, [r1, r6]
	adds	r4, r4, #1
	ldrh	ip, [r0, r6]
	rsb	r5, r4, #200
	mul	r7, ip, r7
	strh	r7, [r2, r6]	@ movhi
	adds	r6, r6, #2
	cmp	r3, r4
	bhi	.L3
.L2:
	rsb	r6, r3, #200
	lsr	r8, r6, #3
	str	r6, [sp, #4]
	lsl	fp, r8, #3
	cmp	r8, #0
	beq	.L4
	add	r3, r0, r3, lsl #1
	movs	r6, #0
	rsb	sl, r0, r1
	rsb	r9, r0, r2
	mov	ip, r4
.L5:
	add	r7, r3, sl
	vldmia	r3, {d16-d17}
	adds	r6, r6, #1
	add	r4, r3, r9
	vld1.16	{q9}, [r7]
	adds	r3, r3, #16
	vmul.i16	q8, q9, q8 @ <= Neon instruction
	cmp	r6, r8
	vst1.16	{q8}, [r4]
	bcc	.L5
	ldr	r3, [sp, #4]
	add	r4, ip, fp
	rsb	r5, fp, r5
	cmp	r3, fp
	beq	.L1
.L4:
	lsls	r4, r4, #1
	movs	r3, #0
	adds	r2, r2, r4
	adds	r1, r1, r4
	adds	r0, r0, r4
.L7:
	ldrh	r4, [r1, r3]
	ldrh	r6, [r0, r3]
	mul	r4, r6, r4
	strh	r4, [r2, r3]	@ movhi
	adds	r3, r3, #2
	subs	r5, r5, #1
	bne	.L7
.L1:
	add	sp, sp, #8
	pop	{r4, r5, r6, r7, r8, r9, sl, fp}
	bx	lr
.L8:
	movs	r5, #200
	mov	r4, r3
	b	.L2
	.size	NeonTest, .-NeonTest
	.ident	"GCC: (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3"
	.section	.note.GNU-stack,"",%progbits