Compiler output using `rustc -Copt-level=3 -Ctarget-cpu=native --emit=asm float_fast.rs -Cllvm-args=-x86-asm-syntax=intel`

float_fast.s

	.text
	.intel_syntax noprefix
	.file	"float_fast.cgu-0.rs"
	.section	.text.dot3_fast,"ax",@progbits
	.globl	dot3_fast
	.p2align	4, 0x90
	.type	dot3_fast,@function
dot3_fast:
	.cfi_startproc
	cmp	rcx, r9
	cmovbe	r9, rcx
	cmp	r9, rsi
	cmovae	r9, rsi
	vxorps	xmm0, xmm0, xmm0
	test	r9, r9
	je	.LBB0_7
	vxorps	xmm0, xmm0, xmm0
	xor	eax, eax
	cmp	r9, 31
	jbe	.LBB0_6
	xor	eax, eax
	mov	r10, r9
	and	r10, -32
	je	.LBB0_6
	lea	rcx, [rdi + 96]
	lea	rsi, [r8 + 96]
	lea	rax, [rdx + 96]
	vxorps	ymm0, ymm0, ymm0
	mov	r11, r10
	vxorps	ymm1, ymm1, ymm1
	vxorps	ymm2, ymm2, ymm2
	vxorps	ymm3, ymm3, ymm3
	.p2align	4, 0x90
.LBB0_4:
	vmovups	xmm4, xmmword ptr [rcx - 96]
	vmovups	xmm5, xmmword ptr [rcx - 64]
	vmovups	xmm6, xmmword ptr [rcx - 32]
	vmovups	xmm7, xmmword ptr [rcx]
	vinsertf128	ymm4, ymm4, xmmword ptr [rcx - 80], 1
	vinsertf128	ymm5, ymm5, xmmword ptr [rcx - 48], 1
	vinsertf128	ymm6, ymm6, xmmword ptr [rcx - 16], 1
	vinsertf128	ymm7, ymm7, xmmword ptr [rcx + 16], 1
	vmovups	xmm8, xmmword ptr [rax - 96]
	vmovups	xmm9, xmmword ptr [rax - 64]
	vmovups	xmm10, xmmword ptr [rax - 32]
	vmovups	xmm11, xmmword ptr [rax]
	vinsertf128	ymm8, ymm8, xmmword ptr [rax - 80], 1
	vinsertf128	ymm9, ymm9, xmmword ptr [rax - 48], 1
	vinsertf128	ymm10, ymm10, xmmword ptr [rax - 16], 1
	vinsertf128	ymm11, ymm11, xmmword ptr [rax + 16], 1
	vmovups	xmm12, xmmword ptr [rsi - 96]
	vmovups	xmm13, xmmword ptr [rsi - 64]
	vmovups	xmm14, xmmword ptr [rsi - 32]
	vmovups	xmm15, xmmword ptr [rsi]
	vinsertf128	ymm12, ymm12, xmmword ptr [rsi - 80], 1
	vinsertf128	ymm13, ymm13, xmmword ptr [rsi - 48], 1
	vinsertf128	ymm14, ymm14, xmmword ptr [rsi - 16], 1
	vinsertf128	ymm15, ymm15, xmmword ptr [rsi + 16], 1
	vmulps	ymm4, ymm4, ymm4
	vmulps	ymm5, ymm5, ymm5
	vmulps	ymm6, ymm6, ymm6
	vmulps	ymm7, ymm7, ymm7
	vmulps	ymm8, ymm8, ymm8
	vmulps	ymm9, ymm9, ymm9
	vmulps	ymm10, ymm10, ymm10
	vmulps	ymm11, ymm11, ymm11
	vmulps	ymm12, ymm12, ymm12
	vmulps	ymm13, ymm13, ymm13
	vmulps	ymm14, ymm14, ymm14
	vmulps	ymm15, ymm15, ymm15
	vaddps	ymm0, ymm4, ymm0
	vaddps	ymm1, ymm5, ymm1
	vaddps	ymm2, ymm6, ymm2
	vaddps	ymm3, ymm7, ymm3
	vaddps	ymm0, ymm0, ymm8
	vaddps	ymm1, ymm1, ymm9
	vaddps	ymm2, ymm2, ymm10
	vaddps	ymm3, ymm3, ymm11
	vaddps	ymm0, ymm0, ymm12
	vaddps	ymm1, ymm1, ymm13
	vaddps	ymm2, ymm2, ymm14
	vaddps	ymm3, ymm3, ymm15
	sub	rcx, -128
	sub	rsi, -128
	sub	rax, -128
	add	r11, -32
	jne	.LBB0_4
	vaddps	ymm0, ymm1, ymm0
	vaddps	ymm0, ymm2, ymm0
	vaddps	ymm0, ymm3, ymm0
	vextractf128	xmm1, ymm0, 1
	vaddps	ymm0, ymm0, ymm1
	vpermilpd	xmm1, xmm0, 1
	vaddps	ymm0, ymm0, ymm1
	vhaddps	ymm0, ymm0, ymm0
	mov	rax, r10
	cmp	r9, r10
	je	.LBB0_7
	.p2align	4, 0x90
.LBB0_6:
	vmovss	xmm1, dword ptr [rdi + 4*rax]
	vmovss	xmm2, dword ptr [rdx + 4*rax]
	vmovss	xmm3, dword ptr [r8 + 4*rax]
	inc	rax
	vmulss	xmm1, xmm1, xmm1
	vmulss	xmm2, xmm2, xmm2
	vmulss	xmm3, xmm3, xmm3
	vaddss	xmm0, xmm1, xmm0
	vaddss	xmm0, xmm0, xmm2
	vaddss	xmm0, xmm0, xmm3
	cmp	rax, r9
	jb	.LBB0_6
.LBB0_7:
	vzeroupper
	ret
.Lfunc_end0:
	.size	dot3_fast, .Lfunc_end0-dot3_fast
	.cfi_endproc


	.section	".note.GNU-stack","",@progbits