clang's ARM NEON codegen test.

gistfile1.txt

void
clang_test(float* out, float* in)
{
  for (int i = 0; i < 4; i++) {
    out[i] = in[i] * in[i];
  }
}

// asm output

LBB0_1:                                 @ =>This Inner Loop Header: Depth=1
  @DEBUG_VALUE: clang_test:out <- R0+0
	@DEBUG_VALUE: clang_test:in <- R1+0
	@DEBUG_VALUE: i <- 0+0
	.loc	1 16 17 prologue_end    @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:16:17
	adds	r3, r1, r2
	.loc	1 17 5                  @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:17:5
Ltmp1:
	vldr	s0, [r3]
Ltmp2:
	.loc	1 16 17                 @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:16:17
	adds	r3, r0, r2
	adds	r2, #4
	.loc	1 17 5                  @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:17:5
Ltmp3:
	vmul.f32	d0, d0, d0
Ltmp4:
	.loc	1 16 17                 @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:16:17
	cmp	r2, #16
	.loc	1 17 5                  @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:17:5
Ltmp5:
	vstr	s0, [r3]
	vorr	d0, d0, d0              @ vnop
Ltmp6:
	.loc	1 16 17                 @ /Users/syoyo/work/ios_dev/clantest/clantest/clangtest.cc:16:17
	bne	LBB0_1