OV - attention-avx512 - Loop 41

	Loop Id: 41	Module: attention-avx512	Source: attention.cpp:43-284 [...]	Coverage: 0.10%

0x6d30 MOV	0x380(%RSP),%RCX

0x6d38 MOV	0x2a8(%RSP),%R8

0x6d40 MOV	0x330(%RSP),%RDI

0x6d48 MOV	0x220(%RSP),%RDX

0x6d50 MOV	0x238(%RSP),%R13

0x6d58 INC	%RCX

0x6d5b ADD	%R8,%RDI

0x6d5e ADD	%R8,%RSI

0x6d61 ADD	%R8,%RDX

0x6d64 VMOVSS	%XMM2,(%R13,%RAX,4)

0x6d6b MOV	%R14,%R8

0x6d6e MOV	%RDX,%RAX

0x6d71 CMP	0x38(%RSP),%R14

0x6d76 MOV	%RCX,%R14

0x6d79 JE	6ce0

0x6d7f MOV	%R14,%R9

0x6d82 MOV	%R14,%R10

0x6d85 AND	$-0x8,%R9

0x6d89 AND	$-0x40,%R10

0x6d8d MOV	%RAX,0x220(%RSP)

0x6d95 CMP	$0x8,%R14

0x6d99 JAE	6db0

0x6d9b VMOVSS	-0x5ed3(%RIP),%XMM1

0x6da3 XOR	%EAX,%EAX

0x6da5 JMP	6ec0

0x6db0 CMP	$0x40,%R14

0x6db4 JAE	6dd0

0x6db6 VMOVSS	-0x5eee(%RIP),%XMM1

0x6dbe XOR	%EAX,%EAX

0x6dc0 JMP	6e6d

0x6dd0 VBROADCASTSS	-0x5f0a(%RIP),%ZMM0

0x6dda MOV	$0x7ffffffffffffff8,%RAX

0x6de4 XOR	%ECX,%ECX

0x6de6 ADD	$-0x38,%RAX

0x6dea AND	%R14,%RAX

0x6ded VMOVAPS	%ZMM0,%ZMM1

0x6df3 VMOVAPS	%ZMM0,%ZMM2

0x6df9 VMOVAPS	%ZMM0,%ZMM3

0x6dff NOP

  (36) 0x6e00 VMAXPS	-0xc0(%RDI,%RCX,4),%ZMM0,%ZMM0

  (36) 0x6e08 VMAXPS	-0x80(%RDI,%RCX,4),%ZMM1,%ZMM1

  (36) 0x6e10 VMAXPS	-0x40(%RDI,%RCX,4),%ZMM2,%ZMM2

  (36) 0x6e18 VMAXPS	(%RDI,%RCX,4),%ZMM3,%ZMM3

  (36) 0x6e1f ADD	$0x40,%RCX

  (36) 0x6e23 CMP	%RCX,%R10

  (36) 0x6e26 JNE	6e00

0x6e28 VMAXPS	%ZMM1,%ZMM0,%ZMM0

0x6e2e VMAXPS	%ZMM3,%ZMM2,%ZMM2

0x6e34 VMAXPS	%ZMM2,%ZMM0,%ZMM0

0x6e3a VEXTRACTF64X4	$0x1,%ZMM0,%YMM1

0x6e41 VMAXPS	%ZMM1,%ZMM0,%ZMM0

0x6e47 VEXTRACTF128	$0x1,%YMM0,%XMM1

0x6e4d VMAXPS	%XMM1,%XMM0,%XMM0

0x6e51 VSHUFPD	$0x1,%XMM0,%XMM0,%XMM1

0x6e56 VMAXPS	%XMM1,%XMM0,%XMM0

0x6e5a VMOVSHDUP	%XMM0,%XMM1

0x6e5e VMAXSS	%XMM1,%XMM0,%XMM1

0x6e62 CMP	%RAX,%R14

0x6e65 JE	6ecd

0x6e67 TEST	$0x38,%R14B

0x6e6b JE	6ec0

0x6e6d MOV	$0x7ffffffffffffff8,%RDX

0x6e77 VBROADCASTSS	%XMM1,%YMM0

0x6e7c MOV	%RAX,%RCX

0x6e7f MOV	%R14,%RAX

0x6e82 AND	%RDX,%RAX

0x6e85 NOPW	%CS:(%RAX,%RAX,1)

  (46) 0x6e90 VMAXPS	(%RSI,%RCX,4),%YMM0,%YMM0

  (46) 0x6e95 ADD	$0x8,%RCX

  (46) 0x6e99 CMP	%RCX,%R9

  (46) 0x6e9c JNE	6e90

0x6e9e VEXTRACTF128	$0x1,%YMM0,%XMM1

0x6ea4 VMAXPS	%XMM1,%XMM0,%XMM0

0x6ea8 VSHUFPD	$0x1,%XMM0,%XMM0,%XMM1

0x6ead VMAXPS	%XMM1,%XMM0,%XMM0

0x6eb1 VMOVSHDUP	%XMM0,%XMM1

0x6eb5 VMAXSS	%XMM1,%XMM0,%XMM1

0x6eb9 JMP	6ec8

  (45) 0x6ec0 VMAXSS	(%RSI,%RAX,4),%XMM1,%XMM1

  (45) 0x6ec5 INC	%RAX

  (45) 0x6ec8 CMP	%RAX,%R14

  (45) 0x6ecb JNE	6ec0

0x6ecd MOV	%R14,0x380(%RSP)

0x6ed5 MOV	%RSI,0x28(%RSP)

0x6eda MOV	%R8,0x328(%RSP)

0x6ee2 VMOVAPS	%XMM1,0x2c0(%RSP)

0x6eeb MOV	%RDI,0x330(%RSP)

0x6ef3 MOV	%R9,0x338(%RSP)

0x6efb CMP	$0x8,%R14

0x6eff JAE	6f10

0x6f01 VXORPS	%XMM2,%XMM2,%XMM2

0x6f05 XOR	%R13D,%R13D

0x6f08 JMP	7c20

0x6f10 CMP	$0x40,%R14

0x6f14 JAE	6f30

0x6f16 VXORPS	%XMM2,%XMM2,%XMM2

0x6f1a XOR	%R13D,%R13D

0x6f1d JMP	7a5a

0x6f30 MOV	$0x7ffffffffffffff8,%RAX

0x6f3a VBROADCASTSS	%XMM1,%ZMM0

0x6f40 VXORPS	%XMM1,%XMM1,%XMM1

0x6f44 XOR	%R13D,%R13D

0x6f47 MOV	%R10,0x3e8(%RSP)

0x6f4f ADD	$-0x38,%RAX

0x6f53 AND	%R14,%RAX

0x6f56 VMOVAPS	%ZMM0,0x480(%RSP)

0x6f5e VXORPS	%XMM0,%XMM0,%XMM0

0x6f62 MOV	%RAX,0x2b0(%RSP)

0x6f6a VMOVAPS	%ZMM0,0x340(%RSP)

0x6f72 VMOVAPS	%ZMM0,0x400(%RSP)

0x6f7a VMOVAPS	%ZMM0,0x440(%RSP)

0x6f82 NOPW	%CS:(%RAX,%RAX,1)

  (37) 0x6f90 VMOVUPS	-0xc0(%RDI,%R13,4),%ZMM0

  (37) 0x6f98 VMOVAPS	0x480(%RSP),%ZMM4

  (37) 0x6fa0 VMOVUPS	(%RDI,%R13,4),%ZMM3

  (37) 0x6fa7 VMOVAPS	%ZMM1,0x4c0(%RSP)

  (37) 0x6faf VMOVUPS	-0x80(%RDI,%R13,4),%ZMM1

  (37) 0x6fb7 VMOVUPS	-0x40(%RDI,%R13,4),%ZMM2

  (37) 0x6fbf VSUBPS	%ZMM4,%ZMM0,%ZMM5

  (37) 0x6fc5 VSUBPS	%ZMM4,%ZMM3,%ZMM0

  (37) 0x6fcb VSUBPS	%ZMM4,%ZMM1,%ZMM6

  (37) 0x6fd1 VSUBPS	%ZMM4,%ZMM2,%ZMM1

  (37) 0x6fd7 VMOVAPS	%ZMM0,0x240(%RSP)

  (37) 0x6fdf VEXTRACTF32X4	$0x3,%ZMM5,%XMM0

  (37) 0x6fe6 VMOVAPS	%ZMM5,0x40(%RSP)

  (37) 0x6fee VMOVAPS	%ZMM6,0x100(%RSP)

  (37) 0x6ff6 VMOVAPS	%ZMM1,0x1c0(%RSP)

  (37) 0x6ffe VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x7007 VZEROUPPER

  (37) 0x700a CALL	89b0 <@plt_start@+0x20>

  (37) 0x700f VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x7018 VMOVSHDUP	0xc0(%RSP),%XMM0

  (37) 0x7021 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7026 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x702f VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7035 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x703e VPERMILPD	$0x1,0xc0(%RSP),%XMM0

  (37) 0x7049 CALL	89b0 <@plt_start@+0x20>

  (37) 0x704e VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x7057 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x705d VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x7066 VPERMILPS	$-0x1,0xc0(%RSP),%XMM0

  (37) 0x7071 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7076 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x707f VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7085 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x708e VMOVAPS	0x40(%RSP),%ZMM0

  (37) 0x7096 VEXTRACTF32X4	$0x2,%ZMM0,%XMM0

  (37) 0x709d VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x70a6 VZEROUPPER

  (37) 0x70a9 CALL	89b0 <@plt_start@+0x20>

  (37) 0x70ae VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x70b7 VMOVSHDUP	0xc0(%RSP),%XMM0

  (37) 0x70c0 CALL	89b0 <@plt_start@+0x20>

  (37) 0x70c5 VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x70ce VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x70d4 VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x70dd VPERMILPD	$0x1,0xc0(%RSP),%XMM0

  (37) 0x70e8 CALL	89b0 <@plt_start@+0x20>

  (37) 0x70ed VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x70f6 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x70fc VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x7105 VPERMILPS	$-0x1,0xc0(%RSP),%XMM0

  (37) 0x7110 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7115 VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x711e VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7124 VINSERTF128	$0x1,0x180(%RSP),%YMM0,%YMM0

  (37) 0x712f VMOVAPS	%YMM0,0x180(%RSP)

  (37) 0x7138 VMOVAPS	0x40(%RSP),%ZMM0

  (37) 0x7140 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (37) 0x7146 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x714f VZEROUPPER

  (37) 0x7152 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7157 VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x7160 VMOVSHDUP	0xc0(%RSP),%XMM0

  (37) 0x7169 CALL	89b0 <@plt_start@+0x20>

  (37) 0x716e VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x7177 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x717d VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x7186 VPERMILPD	$0x1,0xc0(%RSP),%XMM0

  (37) 0x7191 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7196 VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x719f VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x71a5 VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x71ae VPERMILPS	$-0x1,0xc0(%RSP),%XMM0

  (37) 0x71b9 CALL	89b0 <@plt_start@+0x20>

  (37) 0x71be VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x71c7 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x71cd VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x71d6 VMOVAPS	0x40(%RSP),%ZMM0

  (37) 0x71de VZEROUPPER

  (37) 0x71e1 CALL	89b0 <@plt_start@+0x20>

  (37) 0x71e6 VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x71ef VMOVSHDUP	0x40(%RSP),%XMM0

  (37) 0x71f5 CALL	89b0 <@plt_start@+0x20>

  (37) 0x71fa VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x7203 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7209 VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x7212 VPERMILPD	$0x1,0x40(%RSP),%XMM0

  (37) 0x721a CALL	89b0 <@plt_start@+0x20>

  (37) 0x721f VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x7228 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x722e VMOVAPS	%XMM0,0x160(%RSP)

  (37) 0x7237 VPERMILPS	$-0x1,0x40(%RSP),%XMM0

  (37) 0x723f CALL	89b0 <@plt_start@+0x20>

  (37) 0x7244 VMOVAPS	0x160(%RSP),%XMM1

  (37) 0x724d VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7253 VINSERTF128	$0x1,0xc0(%RSP),%YMM0,%YMM0

  (37) 0x725e VMOVAPS	0x340(%RSP),%ZMM1

  (37) 0x7266 VINSERTF64X4	$0x1,0x180(%RSP),%ZMM0,%ZMM0

  (37) 0x726f VADDPS	%ZMM1,%ZMM0,%ZMM1

  (37) 0x7275 VMOVAPS	0x100(%RSP),%ZMM0

  (37) 0x727d VMOVAPS	%ZMM1,0x340(%RSP)

  (37) 0x7285 VEXTRACTF32X4	$0x3,%ZMM0,%XMM0

  (37) 0x728c VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7292 VZEROUPPER

  (37) 0x7295 CALL	89b0 <@plt_start@+0x20>

  (37) 0x729a VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x72a3 VMOVSHDUP	0x40(%RSP),%XMM0

  (37) 0x72a9 CALL	89b0 <@plt_start@+0x20>

  (37) 0x72ae VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x72b7 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x72bd VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x72c6 VPERMILPD	$0x1,0x40(%RSP),%XMM0

  (37) 0x72ce CALL	89b0 <@plt_start@+0x20>

  (37) 0x72d3 VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x72dc VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x72e2 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x72eb VPERMILPS	$-0x1,0x40(%RSP),%XMM0

  (37) 0x72f3 CALL	89b0 <@plt_start@+0x20>

  (37) 0x72f8 VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x7301 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7307 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x7310 VMOVAPS	0x100(%RSP),%ZMM0

  (37) 0x7318 VEXTRACTF32X4	$0x2,%ZMM0,%XMM0

  (37) 0x731f VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7325 VZEROUPPER

  (37) 0x7328 CALL	89b0 <@plt_start@+0x20>

  (37) 0x732d VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x7336 VMOVSHDUP	0x40(%RSP),%XMM0

  (37) 0x733c CALL	89b0 <@plt_start@+0x20>

  (37) 0x7341 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x734a VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7350 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x7359 VPERMILPD	$0x1,0x40(%RSP),%XMM0

  (37) 0x7361 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7366 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x736f VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x7375 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x737e VPERMILPS	$-0x1,0x40(%RSP),%XMM0

  (37) 0x7386 CALL	89b0 <@plt_start@+0x20>

  (37) 0x738b VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x7394 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x739a VINSERTF128	$0x1,0xc0(%RSP),%YMM0,%YMM0

  (37) 0x73a5 VMOVAPS	%YMM0,0xc0(%RSP)

  (37) 0x73ae VMOVAPS	0x100(%RSP),%ZMM0

  (37) 0x73b6 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (37) 0x73bc VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x73c2 VZEROUPPER

  (37) 0x73c5 CALL	89b0 <@plt_start@+0x20>

  (37) 0x73ca VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x73d3 VMOVSHDUP	0x40(%RSP),%XMM0

  (37) 0x73d9 CALL	89b0 <@plt_start@+0x20>

  (37) 0x73de VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x73e7 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x73ed VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x73f6 VPERMILPD	$0x1,0x40(%RSP),%XMM0

  (37) 0x73fe CALL	89b0 <@plt_start@+0x20>

  (37) 0x7403 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x740c VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x7412 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x741b VPERMILPS	$-0x1,0x40(%RSP),%XMM0

  (37) 0x7423 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7428 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x7431 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7437 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x743d VMOVAPS	0x100(%RSP),%ZMM0

  (37) 0x7445 VZEROUPPER

  (37) 0x7448 CALL	89b0 <@plt_start@+0x20>

  (37) 0x744d VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x7456 VMOVSHDUP	0x100(%RSP),%XMM0

  (37) 0x745f CALL	89b0 <@plt_start@+0x20>

  (37) 0x7464 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x746d VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7473 VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x747c VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (37) 0x7487 CALL	89b0 <@plt_start@+0x20>

  (37) 0x748c VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x7495 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x749b VMOVAPS	%XMM0,0x180(%RSP)

  (37) 0x74a4 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (37) 0x74af CALL	89b0 <@plt_start@+0x20>

  (37) 0x74b4 VMOVAPS	0x180(%RSP),%XMM1

  (37) 0x74bd VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x74c3 VINSERTF128	$0x1,0x40(%RSP),%YMM0,%YMM0

  (37) 0x74cb VMOVAPS	0x400(%RSP),%ZMM1

  (37) 0x74d3 VINSERTF64X4	$0x1,0xc0(%RSP),%ZMM0,%ZMM0

  (37) 0x74dc VADDPS	%ZMM1,%ZMM0,%ZMM1

  (37) 0x74e2 VMOVAPS	0x1c0(%RSP),%ZMM0

  (37) 0x74ea VMOVAPS	%ZMM1,0x400(%RSP)

  (37) 0x74f2 VEXTRACTF32X4	$0x3,%ZMM0,%XMM0

  (37) 0x74f9 VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x7502 VZEROUPPER

  (37) 0x7505 CALL	89b0 <@plt_start@+0x20>

  (37) 0x750a VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7510 VMOVSHDUP	0x100(%RSP),%XMM0

  (37) 0x7519 CALL	89b0 <@plt_start@+0x20>

  (37) 0x751e VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7524 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x752a VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7530 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (37) 0x753b CALL	89b0 <@plt_start@+0x20>

  (37) 0x7540 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7546 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x754c VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7552 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (37) 0x755d CALL	89b0 <@plt_start@+0x20>

  (37) 0x7562 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7568 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x756e VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7574 VMOVAPS	0x1c0(%RSP),%ZMM0

  (37) 0x757c VEXTRACTF32X4	$0x2,%ZMM0,%XMM0

  (37) 0x7583 VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x758c VZEROUPPER

  (37) 0x758f CALL	89b0 <@plt_start@+0x20>

  (37) 0x7594 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x759d VMOVSHDUP	0x100(%RSP),%XMM0

  (37) 0x75a6 CALL	89b0 <@plt_start@+0x20>

  (37) 0x75ab VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x75b4 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x75ba VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x75c3 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (37) 0x75ce CALL	89b0 <@plt_start@+0x20>

  (37) 0x75d3 VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x75dc VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x75e2 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x75eb VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (37) 0x75f6 CALL	89b0 <@plt_start@+0x20>

  (37) 0x75fb VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x7604 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x760a VINSERTF128	$0x1,0x40(%RSP),%YMM0,%YMM0

  (37) 0x7612 VMOVAPS	%YMM0,0x40(%RSP)

  (37) 0x7618 VMOVAPS	0x1c0(%RSP),%ZMM0

  (37) 0x7620 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (37) 0x7626 VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x762f VZEROUPPER

  (37) 0x7632 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7637 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x7640 VMOVSHDUP	0x100(%RSP),%XMM0

  (37) 0x7649 CALL	89b0 <@plt_start@+0x20>

  (37) 0x764e VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x7657 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x765d VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x7666 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (37) 0x7671 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7676 VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x767f VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x7685 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x768e VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (37) 0x7699 CALL	89b0 <@plt_start@+0x20>

  (37) 0x769e VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x76a7 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x76ad VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x76b6 VMOVAPS	0x1c0(%RSP),%ZMM0

  (37) 0x76be VZEROUPPER

  (37) 0x76c1 CALL	89b0 <@plt_start@+0x20>

  (37) 0x76c6 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x76cf VMOVSHDUP	0x1c0(%RSP),%XMM0

  (37) 0x76d8 CALL	89b0 <@plt_start@+0x20>

  (37) 0x76dd VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x76e6 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x76ec VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x76f5 VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (37) 0x7700 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7705 VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x770e VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x7714 VMOVAPS	%XMM0,0xc0(%RSP)

  (37) 0x771d VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (37) 0x7728 CALL	89b0 <@plt_start@+0x20>

  (37) 0x772d VMOVAPS	0xc0(%RSP),%XMM1

  (37) 0x7736 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x773c VINSERTF128	$0x1,0x100(%RSP),%YMM0,%YMM0

  (37) 0x7747 VMOVAPS	0x440(%RSP),%ZMM1

  (37) 0x774f VINSERTF64X4	$0x1,0x40(%RSP),%ZMM0,%ZMM0

  (37) 0x7758 VADDPS	%ZMM1,%ZMM0,%ZMM1

  (37) 0x775e VMOVAPS	0x240(%RSP),%ZMM0

  (37) 0x7766 VMOVAPS	%ZMM1,0x440(%RSP)

  (37) 0x776e VEXTRACTF32X4	$0x3,%ZMM0,%XMM0

  (37) 0x7775 VMOVAPS	%XMM0,0x1c0(%RSP)

  (37) 0x777e VZEROUPPER

  (37) 0x7781 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7786 VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x778f VMOVSHDUP	0x1c0(%RSP),%XMM0

  (37) 0x7798 CALL	89b0 <@plt_start@+0x20>

  (37) 0x779d VMOVAPS	0x100(%RSP),%XMM1

  (37) 0x77a6 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x77ac VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x77b5 VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (37) 0x77c0 CALL	89b0 <@plt_start@+0x20>

  (37) 0x77c5 VMOVAPS	0x100(%RSP),%XMM1

  (37) 0x77ce VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x77d4 VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x77dd VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (37) 0x77e8 CALL	89b0 <@plt_start@+0x20>

  (37) 0x77ed VMOVAPS	0x100(%RSP),%XMM1

  (37) 0x77f6 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x77fc VMOVAPS	%XMM0,0x100(%RSP)

  (37) 0x7805 VMOVAPS	0x240(%RSP),%ZMM0

  (37) 0x780d VEXTRACTF32X4	$0x2,%ZMM0,%XMM0

  (37) 0x7814 VMOVAPS	%XMM0,0x1c0(%RSP)

  (37) 0x781d VZEROUPPER

  (37) 0x7820 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7825 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x782b VMOVSHDUP	0x1c0(%RSP),%XMM0

  (37) 0x7834 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7839 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x783f VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7845 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x784b VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (37) 0x7856 CALL	89b0 <@plt_start@+0x20>

  (37) 0x785b VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7861 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x7867 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x786d VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (37) 0x7878 CALL	89b0 <@plt_start@+0x20>

  (37) 0x787d VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7883 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7889 VINSERTF128	$0x1,0x100(%RSP),%YMM0,%YMM0

  (37) 0x7894 VMOVAPS	%YMM0,0x100(%RSP)

  (37) 0x789d VMOVAPS	0x240(%RSP),%ZMM0

  (37) 0x78a5 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (37) 0x78ab VMOVAPS	%XMM0,0x1c0(%RSP)

  (37) 0x78b4 VZEROUPPER

  (37) 0x78b7 CALL	89b0 <@plt_start@+0x20>

  (37) 0x78bc VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x78c2 VMOVSHDUP	0x1c0(%RSP),%XMM0

  (37) 0x78cb CALL	89b0 <@plt_start@+0x20>

  (37) 0x78d0 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x78d6 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x78dc VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x78e2 VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (37) 0x78ed CALL	89b0 <@plt_start@+0x20>

  (37) 0x78f2 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x78f8 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x78fe VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7904 VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (37) 0x790f CALL	89b0 <@plt_start@+0x20>

  (37) 0x7914 VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x791a VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (37) 0x7920 VMOVAPS	%XMM0,0x1c0(%RSP)

  (37) 0x7929 VMOVAPS	0x240(%RSP),%ZMM0

  (37) 0x7931 VZEROUPPER

  (37) 0x7934 CALL	89b0 <@plt_start@+0x20>

  (37) 0x7939 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x793f VMOVSHDUP	0x240(%RSP),%XMM0

  (37) 0x7948 CALL	89b0 <@plt_start@+0x20>

  (37) 0x794d VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7953 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (37) 0x7959 VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x795f VPERMILPD	$0x1,0x240(%RSP),%XMM0

  (37) 0x796a CALL	89b0 <@plt_start@+0x20>

  (37) 0x796f VMOVAPS	0x40(%RSP),%XMM1

  (37) 0x7975 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (37) 0x797b VMOVAPS	%XMM0,0x40(%RSP)

  (37) 0x7981 VPERMILPS	$-0x1,0x240(%RSP),%XMM0

  (37) 0x798c CALL	89b0 <@plt_start@+0x20>

  (37) 0x7991 VMOVAPS	0x40(%RSP),%XMM2

  (37) 0x7997 VMOVAPS	0x4c0(%RSP),%ZMM1

  (37) 0x799f MOV	0x3e8(%RSP),%R10

  (37) 0x79a7 MOV	0x330(%RSP),%RDI

  (37) 0x79af ADD	$0x40,%R13

  (37) 0x79b3 VINSERTPS	$0x30,%XMM0,%XMM2,%XMM0

  (37) 0x79b9 VINSERTF128	$0x1,0x1c0(%RSP),%YMM0,%YMM0

  (37) 0x79c4 VINSERTF64X4	$0x1,0x100(%RSP),%ZMM0,%ZMM0

  (37) 0x79cd VADDPS	%ZMM1,%ZMM0,%ZMM1

  (37) 0x79d3 CMP	%R13,%R10

  (37) 0x79d6 JNE	6f90

0x79dc VMOVAPS	0x400(%RSP),%ZMM0

0x79e4 MOV	0x2b0(%RSP),%R13

0x79ec VADDPS	0x340(%RSP),%ZMM0,%ZMM0

0x79f4 VADDPS	0x440(%RSP),%ZMM0,%ZMM0

0x79fc VADDPS	%ZMM0,%ZMM1,%ZMM0

0x7a02 VEXTRACTF64X4	$0x1,%ZMM0,%YMM1

0x7a09 VADDPS	%ZMM1,%ZMM0,%ZMM0

0x7a0f VEXTRACTF128	$0x1,%YMM0,%XMM1

0x7a15 VADDPS	%XMM1,%XMM0,%XMM0

0x7a19 VSHUFPD	$0x1,%XMM0,%XMM0,%XMM1

0x7a1e VADDPS	%XMM1,%XMM0,%XMM0

0x7a22 VMOVSHDUP	%XMM0,%XMM1

0x7a26 VADDSS	%XMM1,%XMM0,%XMM2

0x7a2a CMP	%R13,%R14

0x7a2d JNE	7a42

0x7a2f VMOVAPS	0x2c0(%RSP),%XMM1

0x7a38 MOV	0x28(%RSP),%RSI

0x7a3d JMP	7c5e

0x7a42 VMOVAPS	0x2c0(%RSP),%XMM1

0x7a4b MOV	0x28(%RSP),%RSI

0x7a50 TEST	$0x38,%R14B

0x7a54 JE	7c20

0x7a5a MOV	%R14,%RAX

0x7a5d MOV	%R13,%R14

0x7a60 MOV	0x338(%RSP),%R13

0x7a68 VXORPS	%XMM0,%XMM0,%XMM0

0x7a6c VBLENDPS	$0x1,%XMM2,%XMM0,%XMM2

0x7a72 MOV	%RAX,%RCX

0x7a75 MOV	$0x7ffffffffffffff8,%RAX

0x7a7f VBROADCASTSS	%XMM1,%YMM0

0x7a84 AND	%RAX,%RCX

0x7a87 MOV	%RCX,0x2b0(%RSP)

0x7a8f VMOVAPS	%YMM0,0xc0(%RSP)

0x7a98 NOPL	(%RAX,%RAX,1)

  (44) 0x7aa0 VMOVUPS	(%RSI,%R14,4),%YMM0

  (44) 0x7aa6 VMOVAPS	%YMM2,0x100(%RSP)

  (44) 0x7aaf VSUBPS	0xc0(%RSP),%YMM0,%YMM0

  (44) 0x7ab8 VMOVAPS	%YMM0,0x240(%RSP)

  (44) 0x7ac1 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (44) 0x7ac7 VMOVAPS	%XMM0,0x1c0(%RSP)

  (44) 0x7ad0 VZEROUPPER

  (44) 0x7ad3 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7ad8 VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7ade VMOVSHDUP	0x1c0(%RSP),%XMM0

  (44) 0x7ae7 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7aec VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7af2 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (44) 0x7af8 VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7afe VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (44) 0x7b09 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7b0e VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7b14 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (44) 0x7b1a VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7b20 VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (44) 0x7b2b CALL	89b0 <@plt_start@+0x20>

  (44) 0x7b30 VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7b36 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (44) 0x7b3c VMOVAPS	%XMM0,0x1c0(%RSP)

  (44) 0x7b45 VMOVAPS	0x240(%RSP),%YMM0

  (44) 0x7b4e VZEROUPPER

  (44) 0x7b51 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7b56 VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7b5c VMOVSHDUP	0x240(%RSP),%XMM0

  (44) 0x7b65 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7b6a VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7b70 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (44) 0x7b76 VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7b7c VPERMILPD	$0x1,0x240(%RSP),%XMM0

  (44) 0x7b87 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7b8c VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7b92 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (44) 0x7b98 VMOVAPS	%XMM0,0x40(%RSP)

  (44) 0x7b9e VPERMILPS	$-0x1,0x240(%RSP),%XMM0

  (44) 0x7ba9 CALL	89b0 <@plt_start@+0x20>

  (44) 0x7bae VMOVAPS	0x40(%RSP),%XMM1

  (44) 0x7bb4 VMOVAPS	0x100(%RSP),%YMM2

  (44) 0x7bbd MOV	0x28(%RSP),%RSI

  (44) 0x7bc2 ADD	$0x8,%R14

  (44) 0x7bc6 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (44) 0x7bcc VINSERTF128	$0x1,0x1c0(%RSP),%YMM0,%YMM0

  (44) 0x7bd7 VADDPS	%YMM2,%YMM0,%YMM2

  (44) 0x7bdb CMP	%R14,%R13

  (44) 0x7bde JNE	7aa0

0x7be4 VEXTRACTF128	$0x1,%YMM2,%XMM0

0x7bea MOV	0x380(%RSP),%R14

0x7bf2 MOV	0x2b0(%RSP),%R13

0x7bfa VADDPS	%XMM0,%XMM2,%XMM0

0x7bfe VSHUFPD	$0x1,%XMM0,%XMM0,%XMM1

0x7c03 VADDPS	%XMM1,%XMM0,%XMM0

0x7c07 VMOVSHDUP	%XMM0,%XMM1

0x7c0b VADDSS	%XMM1,%XMM0,%XMM2

0x7c0f VMOVAPS	0x2c0(%RSP),%XMM1

0x7c18 CMP	%R13,%R14

0x7c1b JE	7c5e

0x7c1d NOPL	(%RAX)

  (38) 0x7c20 VMOVSS	(%RSI,%R13,4),%XMM0

  (38) 0x7c26 VMOVAPS	%XMM2,0x240(%RSP)

  (38) 0x7c2f VSUBSS	%XMM1,%XMM0,%XMM0

  (38) 0x7c33 VZEROUPPER

  (38) 0x7c36 CALL	89b0 <@plt_start@+0x20>

  (38) 0x7c3b VMOVAPS	0x240(%RSP),%XMM2

  (38) 0x7c44 VMOVAPS	0x2c0(%RSP),%XMM1

  (38) 0x7c4d MOV	0x28(%RSP),%RSI

  (38) 0x7c52 INC	%R13

  (38) 0x7c55 VADDSS	%XMM2,%XMM0,%XMM2

  (38) 0x7c59 CMP	%R13,%R14

  (38) 0x7c5c JNE	7c20

0x7c5e VMOVAPS	%XMM2,0x240(%RSP)

0x7c67 CMP	$0x8,%R14

0x7c6b JAE	7c80

0x7c6d MOV	0x220(%RSP),%RCX

0x7c75 XOR	%R14D,%R14D

0x7c78 JMP	8160

0x7c80 CMP	$0x10,%R14

0x7c84 JAE	7e40

0x7c8a MOV	%R14,%RAX

0x7c8d XOR	%R14D,%R14D

0x7c90 VBROADCASTSS	%XMM1,%YMM0

0x7c95 MOV	%R14,%R13

0x7c98 MOV	%RAX,%R14

0x7c9b MOV	$0x7ffffffffffffff8,%RAX

0x7ca5 VBROADCASTSS	%XMM2,%YMM1

0x7caa AND	%RAX,%R14

0x7cad VMOVAPS	%YMM0,0xc0(%RSP)

0x7cb6 VMOVAPS	%YMM1,0x340(%RSP)

0x7cbf NOP

  (42) 0x7cc0 VMOVUPS	(%RSI,%R13,4),%YMM0

  (42) 0x7cc6 VSUBPS	0xc0(%RSP),%YMM0,%YMM0

  (42) 0x7ccf VMOVAPS	%YMM0,0x1c0(%RSP)

  (42) 0x7cd8 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (42) 0x7cde VMOVAPS	%XMM0,0x100(%RSP)

  (42) 0x7ce7 VZEROUPPER

  (42) 0x7cea CALL	89b0 <@plt_start@+0x20>

  (42) 0x7cef VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7cf5 VMOVSHDUP	0x100(%RSP),%XMM0

  (42) 0x7cfe CALL	89b0 <@plt_start@+0x20>

  (42) 0x7d03 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7d09 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (42) 0x7d0f VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7d15 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (42) 0x7d20 CALL	89b0 <@plt_start@+0x20>

  (42) 0x7d25 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7d2b VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (42) 0x7d31 VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7d37 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (42) 0x7d42 CALL	89b0 <@plt_start@+0x20>

  (42) 0x7d47 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7d4d VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (42) 0x7d53 VMOVAPS	%XMM0,0x100(%RSP)

  (42) 0x7d5c VMOVAPS	0x1c0(%RSP),%YMM0

  (42) 0x7d65 VZEROUPPER

  (42) 0x7d68 CALL	89b0 <@plt_start@+0x20>

  (42) 0x7d6d VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7d73 VMOVSHDUP	0x1c0(%RSP),%XMM0

  (42) 0x7d7c CALL	89b0 <@plt_start@+0x20>

  (42) 0x7d81 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7d87 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (42) 0x7d8d VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7d93 VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (42) 0x7d9e CALL	89b0 <@plt_start@+0x20>

  (42) 0x7da3 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7da9 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (42) 0x7daf VMOVAPS	%XMM0,0x40(%RSP)

  (42) 0x7db5 VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (42) 0x7dc0 CALL	89b0 <@plt_start@+0x20>

  (42) 0x7dc5 VMOVAPS	0x40(%RSP),%XMM1

  (42) 0x7dcb MOV	0x220(%RSP),%RCX

  (42) 0x7dd3 MOV	0x28(%RSP),%RSI

  (42) 0x7dd8 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (42) 0x7dde VINSERTF128	$0x1,0x100(%RSP),%YMM0,%YMM0

  (42) 0x7de9 VDIVPS	0x340(%RSP),%YMM0,%YMM0

  (42) 0x7df2 VMOVUPS	%YMM0,(%RCX,%R13,4)

  (42) 0x7df8 ADD	$0x8,%R13

  (42) 0x7dfc CMP	%R13,0x338(%RSP)

  (42) 0x7e04 JNE	7cc0

0x7e0a VMOVAPS	0x2c0(%RSP),%XMM1

0x7e13 VMOVAPS	0x240(%RSP),%XMM2

0x7e1c MOV	0x380(%RSP),%RAX

0x7e24 CMP	%R14,%RAX

0x7e27 JNE	8160

0x7e2d JMP	81aa

0x7e40 MOV	%R14,%RAX

0x7e43 AND	$-0x10,%RAX

0x7e47 VBROADCASTSS	%XMM1,%ZMM0

0x7e4d VBROADCASTSS	%XMM2,%ZMM1

0x7e53 MOV	%R14,%RCX

0x7e56 XOR	%R13D,%R13D

0x7e59 MOV	%RAX,0x440(%RSP)

0x7e61 MOV	$0x7ffffffffffffff8,%RAX

0x7e6b LEA	-0x8(%RAX),%R14

0x7e6f VMOVAPS	%ZMM0,0x340(%RSP)

0x7e77 VMOVAPS	%ZMM1,0x400(%RSP)

0x7e7f AND	%RCX,%R14

0x7e82 NOPW	%CS:(%RAX,%RAX,1)

  (39) 0x7e90 VMOVUPS	(%RSI,%R13,4),%ZMM0

  (39) 0x7e97 VSUBPS	0x340(%RSP),%ZMM0,%ZMM0

  (39) 0x7e9f VMOVAPS	%ZMM0,0x1c0(%RSP)

  (39) 0x7ea7 VEXTRACTF32X4	$0x3,%ZMM0,%XMM0

  (39) 0x7eae VMOVAPS	%XMM0,0x100(%RSP)

  (39) 0x7eb7 VZEROUPPER

  (39) 0x7eba CALL	89b0 <@plt_start@+0x20>

  (39) 0x7ebf VMOVAPS	%XMM0,0x40(%RSP)

  (39) 0x7ec5 VMOVSHDUP	0x100(%RSP),%XMM0

  (39) 0x7ece CALL	89b0 <@plt_start@+0x20>

  (39) 0x7ed3 VMOVAPS	0x40(%RSP),%XMM1

  (39) 0x7ed9 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (39) 0x7edf VMOVAPS	%XMM0,0x40(%RSP)

  (39) 0x7ee5 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (39) 0x7ef0 CALL	89b0 <@plt_start@+0x20>

  (39) 0x7ef5 VMOVAPS	0x40(%RSP),%XMM1

  (39) 0x7efb VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (39) 0x7f01 VMOVAPS	%XMM0,0x40(%RSP)

  (39) 0x7f07 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (39) 0x7f12 CALL	89b0 <@plt_start@+0x20>

  (39) 0x7f17 VMOVAPS	0x40(%RSP),%XMM1

  (39) 0x7f1d VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (39) 0x7f23 VMOVAPS	%XMM0,0x40(%RSP)

  (39) 0x7f29 VMOVAPS	0x1c0(%RSP),%ZMM0

  (39) 0x7f31 VEXTRACTF32X4	$0x2,%ZMM0,%XMM0

  (39) 0x7f38 VMOVAPS	%XMM0,0x100(%RSP)

  (39) 0x7f41 VZEROUPPER

  (39) 0x7f44 CALL	89b0 <@plt_start@+0x20>

  (39) 0x7f49 VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x7f52 VMOVSHDUP	0x100(%RSP),%XMM0

  (39) 0x7f5b CALL	89b0 <@plt_start@+0x20>

  (39) 0x7f60 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x7f69 VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (39) 0x7f6f VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x7f78 VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (39) 0x7f83 CALL	89b0 <@plt_start@+0x20>

  (39) 0x7f88 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x7f91 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (39) 0x7f97 VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x7fa0 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (39) 0x7fab CALL	89b0 <@plt_start@+0x20>

  (39) 0x7fb0 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x7fb9 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (39) 0x7fbf VINSERTF128	$0x1,0x40(%RSP),%YMM0,%YMM0

  (39) 0x7fc7 VMOVAPS	%YMM0,0x40(%RSP)

  (39) 0x7fcd VMOVAPS	0x1c0(%RSP),%ZMM0

  (39) 0x7fd5 VEXTRACTF128	$0x1,%YMM0,%XMM0

  (39) 0x7fdb VMOVAPS	%XMM0,0x100(%RSP)

  (39) 0x7fe4 VZEROUPPER

  (39) 0x7fe7 CALL	89b0 <@plt_start@+0x20>

  (39) 0x7fec VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x7ff5 VMOVSHDUP	0x100(%RSP),%XMM0

  (39) 0x7ffe CALL	89b0 <@plt_start@+0x20>

  (39) 0x8003 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x800c VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (39) 0x8012 VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x801b VPERMILPD	$0x1,0x100(%RSP),%XMM0

  (39) 0x8026 CALL	89b0 <@plt_start@+0x20>

  (39) 0x802b VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x8034 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (39) 0x803a VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x8043 VPERMILPS	$-0x1,0x100(%RSP),%XMM0

  (39) 0x804e CALL	89b0 <@plt_start@+0x20>

  (39) 0x8053 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x805c VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (39) 0x8062 VMOVAPS	%XMM0,0x100(%RSP)

  (39) 0x806b VMOVAPS	0x1c0(%RSP),%ZMM0

  (39) 0x8073 VZEROUPPER

  (39) 0x8076 CALL	89b0 <@plt_start@+0x20>

  (39) 0x807b VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x8084 VMOVSHDUP	0x1c0(%RSP),%XMM0

  (39) 0x808d CALL	89b0 <@plt_start@+0x20>

  (39) 0x8092 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x809b VINSERTPS	$0x10,%XMM0,%XMM1,%XMM0

  (39) 0x80a1 VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x80aa VPERMILPD	$0x1,0x1c0(%RSP),%XMM0

  (39) 0x80b5 CALL	89b0 <@plt_start@+0x20>

  (39) 0x80ba VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x80c3 VINSERTPS	$0x20,%XMM0,%XMM1,%XMM0

  (39) 0x80c9 VMOVAPS	%XMM0,0xc0(%RSP)

  (39) 0x80d2 VPERMILPS	$-0x1,0x1c0(%RSP),%XMM0

  (39) 0x80dd CALL	89b0 <@plt_start@+0x20>

  (39) 0x80e2 VMOVAPS	0xc0(%RSP),%XMM1

  (39) 0x80eb MOV	0x220(%RSP),%RCX

  (39) 0x80f3 MOV	0x28(%RSP),%RSI

  (39) 0x80f8 VINSERTPS	$0x30,%XMM0,%XMM1,%XMM0

  (39) 0x80fe VINSERTF128	$0x1,0x100(%RSP),%YMM0,%YMM0

  (39) 0x8109 VINSERTF64X4	$0x1,0x40(%RSP),%ZMM0,%ZMM0

  (39) 0x8112 VDIVPS	0x400(%RSP),%ZMM0,%ZMM0

  (39) 0x811a VMOVUPS	%ZMM0,(%RCX,%R13,4)

  (39) 0x8121 ADD	$0x10,%R13

  (39) 0x8125 CMP	%R13,0x440(%RSP)

  (39) 0x812d JNE	7e90

0x8133 VMOVAPS	0x2c0(%RSP),%XMM1

0x813c VMOVAPS	0x240(%RSP),%XMM2

0x8145 MOV	0x380(%RSP),%RAX

0x814d CMP	%R14,%RAX

0x8150 JE	81aa

0x8152 TEST	$0x8,%AL

0x8154 JNE	7c90

0x815a NOPW	(%RAX,%RAX,1)

  (43) 0x8160 VMOVSS	(%RSI,%R14,4),%XMM0

  (43) 0x8166 MOV	%RCX,%R13

  (43) 0x8169 VSUBSS	%XMM1,%XMM0,%XMM0

  (43) 0x816d VZEROUPPER

  (43) 0x8170 CALL	89b0 <@plt_start@+0x20>

  (43) 0x8175 VMOVAPS	0x240(%RSP),%XMM2

  (43) 0x817e VMOVAPS	0x2c0(%RSP),%XMM1

  (43) 0x8187 MOV	0x380(%RSP),%RAX

  (43) 0x818f MOV	0x28(%RSP),%RSI

  (43) 0x8194 MOV	%R13,%RCX

  (43) 0x8197 VDIVSS	%XMM2,%XMM0,%XMM0

  (43) 0x819b VMOVSS	%XMM0,(%R13,%R14,4)

  (43) 0x81a2 INC	%R14

  (43) 0x81a5 CMP	%R14,%RAX

  (43) 0x81a8 JNE	8160

0x81aa MOV	0x328(%RSP),%RAX

0x81b2 LEA	0x1(%RAX),%R14

0x81b6 CMP	0x38(%RSP),%R14

0x81bb JAE	6d30

0x81c1 MOV	%RAX,%RDI

0x81c4 IMUL	0x3e0(%RSP),%RDI

0x81cd MOV	0x3d8(%RSP),%RDX

0x81d5 LEA	(,%RAX,4),%RAX

0x81dd XOR	%ESI,%ESI

0x81df SUB	%RAX,%RDX

0x81e2 ADD	$0x4,%RDI

0x81e6 ADD	0x80(%RSP),%RDI

0x81ee VZEROUPPER

0x81f1 CALL	89c0 <@plt_start@+0x30>

0x81f6 VMOVAPS	0x240(%RSP),%XMM2

0x81ff MOV	0x328(%RSP),%RAX

0x8207 MOV	0x28(%RSP),%RSI

0x820c JMP	6d30

/home/eoseret/llm-attention/attention.cpp: 43 - 284

--------------------------------------------------------------------------------

43:    for (int row = 0; row < N; ++row) {

44:       const float *S_row = &S[row * N];

45:

46:       float max_val = -FLT_MAX;

47:       for (int idx = 0; idx <= row; ++idx) // vectorised

48:          if (S_row[idx] > max_val) max_val = S_row[idx];

49:

50:       float sum = 0.0f;

51: #pragma clang loop vectorize(enable)

52:       for (int idx = 0; idx <= row; ++idx) // vectorised

53:          sum += expf(S_row[idx] - max_val);

54:

55:       for (int idx = 0; idx <= row; ++idx)  //vectorised

56:          P[row * N + idx] = expf(S_row[idx] - max_val) / sum;

57:

58:       for (int idx = row + 1; idx < N; ++idx)

59:          P[row * N + idx] = 0.0f;

60:

61:       D[row] = sum;

[...]

284:    for (size_t r = 0; r < rept; r++)

Coverage (%)	Name	Source Location	Module

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

min	med	avg	max

Percentile Index	10	20	30	40	50	60	70	80	90	100
Value

Path /

Metric	Value
CQA speedup if no scalar integer	2.00
CQA speedup if FP arith vectorized	1.02
CQA speedup if fully vectorized	1.55
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.75
Bottlenecks	micro-operation queue,
Function	main
Source	attention.cpp:43-44,attention.cpp:47-47,attention.cpp:52-52,attention.cpp:55-55,attention.cpp:58-61,attention.cpp:284-284
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	25.75
CQA cycles if no scalar integer	12.88
CQA cycles if FP arith vectorized	25.14
CQA cycles if fully vectorized	16.61
Front-end cycles	25.75
P0 cycles	11.67
P1 cycles	11.67
P2 cycles	11.67
P3 cycles	11.67
P4 cycles	11.67
P5 cycles	11.67
P6 cycles	14.75
P7 cycles	14.75
P8 cycles	14.75
P9 cycles	14.75
P10 cycles	9.00
P11 cycles	11.50
P12 cycles	11.50
P13 cycles	10.00
P14 cycles	9.00
P15 cycles	9.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	207.00
Nb uops	206.00
Nb loads	36.00
Nb stores	22.00
Nb stack references	21.00
FLOP/cycle	3.18
Nb FLOP add-sub	82.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	42.87
Bytes prefetched	0.00
Bytes loaded	508.00
Bytes stored	596.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	46.15
Vectorization ratio load	44.00
Vectorization ratio store	50.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	57.14
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	43.24
Vector-efficiency ratio all	28.32
Vector-efficiency ratio load	26.25
Vector-efficiency ratio store	42.33
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	40.18
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	23.82

Metric	Value
CQA speedup if no scalar integer	2.00
CQA speedup if FP arith vectorized	1.02
CQA speedup if fully vectorized	1.55
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.75
Bottlenecks	micro-operation queue,
Function	main
Source	attention.cpp:43-44,attention.cpp:47-47,attention.cpp:52-52,attention.cpp:55-55,attention.cpp:58-61,attention.cpp:284-284
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	25.75
CQA cycles if no scalar integer	12.88
CQA cycles if FP arith vectorized	25.14
CQA cycles if fully vectorized	16.61
Front-end cycles	25.75
P0 cycles	11.67
P1 cycles	11.67
P2 cycles	11.67
P3 cycles	11.67
P4 cycles	11.67
P5 cycles	11.67
P6 cycles	14.75
P7 cycles	14.75
P8 cycles	14.75
P9 cycles	14.75
P10 cycles	9.00
P11 cycles	11.50
P12 cycles	11.50
P13 cycles	10.00
P14 cycles	9.00
P15 cycles	9.00
DIV/SQRT cycles	0.00
Inter-iter dependencies cycles	NA
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	207.00
Nb uops	206.00
Nb loads	36.00
Nb stores	22.00
Nb stack references	21.00
FLOP/cycle	3.18
Nb FLOP add-sub	82.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	42.87
Bytes prefetched	0.00
Bytes loaded	508.00
Bytes stored	596.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	46.15
Vectorization ratio load	44.00
Vectorization ratio store	50.00
Vectorization ratio mul	NA
Vectorization ratio add_sub	57.14
Vectorization ratio fma	NA
Vectorization ratio div_sqrt	NA
Vectorization ratio other	43.24
Vector-efficiency ratio all	28.32
Vector-efficiency ratio load	26.25
Vector-efficiency ratio store	42.33
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	40.18
Vector-efficiency ratio fma	NA
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	23.82

Path /

Average path: Display a virtual path defined by average values of all real paths

Function	main
Source file and lines	attention.cpp:43-284
Module	attention-avx512

The loop is defined in /home/eoseret/llm-attention/attention.cpp:43-61,284.

Analyzed code is defined in /home/eoseret/llm-attention/attention.cpp:43-44,47,52,55,58-61,284.

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Too many paths. If you really need to analyze all of the 490 paths individually, rerun with max-paths=490
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

Try to simplify control and/or increase the maximum number of paths per function/loop through the 'max-paths-nb' option.

This loop has 490 execution paths.

The presence of multiple execution paths is typically the main/first bottleneck.
Try to simplify control inside loop: ideally, try to remove all conditional expressions, for example by (if applicable):

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Detected a slowdown caused by scalar integer instructions (typically used for address computation). By removing them, you can lower the cost of an iteration from 25.75 to 12.88 cycles (2.00x speedup).

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Your loop is poorly vectorized. Only 28% of vector register length is used (average across all SSE/AVX instructions). By fully vectorizing your loop, you can lower the cost of an iteration from 25.75 to 16.61 cycles (1.55x speedup).

Details

46% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

44% of SSE/AVX loads are used in vector version.
50% of SSE/AVX stores are used in vector version.
57% of SSE/AVX addition or subtraction instructions are used in vector version.
43% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

No data for this section

CALL instructions

Detected function call instructions.

Details

Calling (and then returning from) a function prevents many compiler optimizations (like vectorization), breaks control flow (which reduces pipeline performance) and executes extra instructions to save/restore the registers used inside it, which is very expensive (dozens of cycles). Consider to inline small functions.

unknown: 1 occurrences►
- /home/eoseret/llm-attention/attention.cpp:59

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

VMOVAPS: 6 occurrences►
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:55
- /home/eoseret/llm-attention/attention.cpp:55

Vector unaligned load/store instructions

Detected 6 suboptimal vector unaligned load/store instructions.

Details

VEXTRACTF128: 4 occurrences►
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
VEXTRACTF64X4: 2 occurrences►
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:52

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

4 SSE or AVX instructions are processing arithmetic or math operations on single precision FP elements in scalar mode (one at a time). 8 SSE or AVX instructions are processing arithmetic or math operations on single precision FP elements in vector mode (four at a time). 8 AVX-512 instructions are processing arithmetic or math operations on single precision FP elements in vector mode (sixteen at a time).

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 82 FP arithmetical operations:

82: addition or subtraction

The binary loop is loading 508 bytes (127 single precision FP elements). The binary loop is storing 596 bytes (149 single precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.

General properties

nb instructions	207
nb uops	206
loop length	1121
used x86 registers	11
used mmx registers	0
used xmm registers	3
used ymm registers	3
used zmm registers	4
nb stack references	21

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	25.75 cycles
front end	25.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	11.67	11.67	11.67	11.67	11.67	11.67	14.75	14.75	14.75	14.75	9.00	11.50	11.50	10.00	9.00	9.00
cycles	11.67	11.67	11.67	11.67	11.67	11.67	14.75	14.75	14.75	14.75	9.00	11.50	11.50	10.00	9.00	9.00

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	25.75
Dispatch	14.75
Overall L1	25.75

Vectorization ratios

INT

all	1%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	3%

all	79%
load	78%
store	91%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	80%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	75%

INT+FP

all	46%
load	44%
store	50%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	57%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Vector efficiency ratios

INT

all	12%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

all	40%
load	37%
store	67%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	51%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	33%

INT+FP

all	28%
load	26%
store	42%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	40%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	23%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 25.75 cycles. At this rate:

7% of peak load performance is reached (19.73 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
18% of peak store performance is reached (23.15 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

Performance is limited by instruction throughput (loading/decoding program instructions to execution core) (front-end is a bottleneck). By removing all these bottlenecks, you can lower the cost of an iteration from 25.75 to 14.75 cycles (1.75x speedup).

ASM code

In the binary file, the address of the loop is: 6d7f

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
MOV 0x380(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x2a8(%RSP),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x330(%RSP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x220(%RSP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x238(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
INC %RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD %R8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD %R8,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVSS %XMM2,(%R13,%RAX,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (6.3%)
MOV %R14,%R8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %RDX,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMP 0x38(%RSP),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV %RCX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
JE 6ce0 <main+0x2290>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %R14,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
AND $-0x8,%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
AND $-0x40,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RAX,0x220(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6db0 <main+0x2360>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSS -0x5ed3(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (6.3%)
XOR %EAX,%EAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 6ec0 <main+0x2470>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x40,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6dd0 <main+0x2380>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSS -0x5eee(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (6.3%)
XOR %EAX,%EAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 6e6d <main+0x241d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
VBROADCASTSS -0x5f0a(%RIP),%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (6.3%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %ECX,%ECX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
ADD $-0x38,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
AND %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPS %ZMM0,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPS %ZMM0,%ZMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
VMAXPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VMAXPS %ZMM3,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VMAXPS %ZMM2,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF64X4 $0x1,%ZMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	4	0.50	vect (50.0%)
VMAXPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VMAXSS %XMM1,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
CMP %RAX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 6ecd <main+0x247d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x38,%R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 6ec0 <main+0x2470>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV $0x7ffffffffffffff8,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %RAX,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND %RDX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VMAXSS %XMM1,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
JMP 6ec8 <main+0x2478>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV %R14,0x380(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %RSI,0x28(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %R8,0x328(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %XMM1,0x2c0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
MOV %RDI,0x330(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %R9,0x338(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6f10 <main+0x24c0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VXORPS %XMM2,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 7c20 <main+0x31d0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x40,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6f30 <main+0x24e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VXORPS %XMM2,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 7a5a <main+0x300a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
VXORPS %XMM1,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R10,0x3e8(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
ADD $-0x38,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
AND %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,0x480(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VXORPS %XMM0,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MOV %RAX,0x2b0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %ZMM0,0x340(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM0,0x400(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM0,0x440(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS 0x400(%RSP),%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
MOV 0x2b0(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VADDPS 0x340(%RSP),%ZMM0,%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPS 0x440(%RSP),%ZMM0,%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPS %ZMM0,%ZMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF64X4 $0x1,%ZMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	4	0.50	vect (50.0%)
VADDPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VADDSS %XMM1,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
CMP %R13,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JNE 7a42 <main+0x2ff2>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
JMP 7c5e <main+0x320e>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
TEST $0x38,%R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 7c20 <main+0x31d0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R13,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV 0x338(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VXORPS %XMM0,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VBLENDPS $0x1,%XMM2,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
MOV %RAX,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
AND %RAX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RCX,0x2b0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %YMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VEXTRACTF128 $0x1,%YMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%R14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x2b0(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VADDPS %XMM0,%XMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VADDSS %XMM1,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
CMP %R13,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 7c5e <main+0x320e>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS %XMM2,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 7c80 <main+0x3230>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV 0x220(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
XOR %R14D,%R14D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
JMP 8160 <main+0x3710>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x10,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 7e40 <main+0x33f0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
XOR %R14D,%R14D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %R14,%R13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %RAX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM2,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
AND %RAX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVAPS %YMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
VMOVAPS %YMM1,0x340(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
CMP %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JNE 8160 <main+0x3710>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
JMP 81aa <main+0x375a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x10,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
VBROADCASTSS %XMM2,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %R14,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %RAX,0x440(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
LEA -0x8(%RAX),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,0x340(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM1,0x400(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
AND %RCX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
CMP %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 81aa <main+0x375a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x8,%AL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JNE 7c90 <main+0x3240>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
NOPW (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV 0x328(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA 0x1(%RAX),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP 0x38(%RSP),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 6d30 <main+0x22e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
IMUL 0x3e0(%RSP),%RDI	1	0.33	0.33	0.33	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.33	N/A
MOV 0x3d8(%RSP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
LEA (,%RAX,4),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %ESI,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
SUB %RAX,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD $0x4,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD 0x80(%RSP),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 89c0 <@plt_start@+0x30>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0.50	N/A
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x328(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
JMP 6d30 <main+0x22e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A

Function	main
Source file and lines	attention.cpp:43-284
Module	attention-avx512

The loop is defined in /home/eoseret/llm-attention/attention.cpp:43-61,284.

Analyzed code is defined in /home/eoseret/llm-attention/attention.cpp:43-44,47,52,55,58-61,284.

Warnings:

Non-innermost loop: analyzing only self part (ignoring child loops).
Ignoring paths for analysis
Too many paths. If you really need to analyze all of the 490 paths individually, rerun with max-paths=490
RecMII not computed since number of paths is unknown or > max_paths
Streams not analyzed since number of paths is unknown or > max_paths

hoisting them (moving them outside the loop)
turning them into conditional moves, MIN or MAX

Ex: if (x<0) x=0 => x = (x<0 ? 0 : x) (or MAX(0,x) after defining the corresponding macro)

gain
potential
hint
expert

Code clean check

Workaround

Try to reorganize arrays of structures to structures of arrays
Consider to permute loops (see vectorization gain report)

Vectorization

Details

46% of SSE/AVX instructions are used in vector version (process two or more data elements in vector registers):

44% of SSE/AVX loads are used in vector version.
50% of SSE/AVX stores are used in vector version.
57% of SSE/AVX addition or subtraction instructions are used in vector version.
43% of SSE/AVX instructions that are not load, store, addition, subtraction nor multiply instructions are used in vector version.

Since your execution units are vector units, only a fully vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one:
- recompile with ffast-math (included in Ofast) to extend loop vectorization to FP reductions.
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

Execution units bottlenecks

Found no such bottlenecks but see expert reports for more complex bottlenecks.

No data for this section

CALL instructions

Detected function call instructions.

Details

unknown: 1 occurrences►
- /home/eoseret/llm-attention/attention.cpp:59

Complex instructions

Detected COMPLEX INSTRUCTIONS.

Details

These instructions generate more than one micro-operation and only one of them can be decoded during a cycle and the extra micro-operations increase pressure on execution units.

VMOVAPS: 6 occurrences►
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:55
- /home/eoseret/llm-attention/attention.cpp:55

Vector unaligned load/store instructions

Detected 6 suboptimal vector unaligned load/store instructions.

Details

VEXTRACTF128: 4 occurrences►
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:52
- /home/eoseret/llm-attention/attention.cpp:52
VEXTRACTF64X4: 2 occurrences►
- /home/eoseret/llm-attention/attention.cpp:47
- /home/eoseret/llm-attention/attention.cpp:52

Workaround

Pass to your compiler a micro-architecture specialization option:
- use march=native
Use vector aligned instructions:
1. align your arrays on 64 bytes boundaries: replace { void *p = malloc (size); } with { void *p; posix_memalign (&p, 64, size); }.
2. inform your compiler that your arrays are vector aligned: if array 'foo' is 64 bytes-aligned, define a pointer 'p_foo' as __builtin_assume_aligned (foo, 64) and use it instead of 'foo' in the loop.

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

The binary loop is composed of 82 FP arithmetical operations:

82: addition or subtraction

The binary loop is loading 508 bytes (127 single precision FP elements). The binary loop is storing 596 bytes (149 single precision FP elements).

Arithmetic intensity

Arithmetic intensity is 0.07 FP operations per loaded or stored byte.

General properties

nb instructions	207
nb uops	206
loop length	1121
used x86 registers	11
used mmx registers	0
used xmm registers	3
used ymm registers	3
used zmm registers	4
nb stack references	21

Front-end

MACRO FUSION NOT POSSIBLE FIT IN UOP CACHE

micro-operation queue	25.75 cycles
front end	25.75 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15
uops	11.67	11.67	11.67	11.67	11.67	11.67	14.75	14.75	14.75	14.75	9.00	11.50	11.50	10.00	9.00	9.00
cycles	11.67	11.67	11.67	11.67	11.67	11.67	14.75	14.75	14.75	14.75	9.00	11.50	11.50	10.00	9.00	9.00

Execution ports to units layout:

P0: ALU
P1: ALU
P2: ALU
P3: ALU, BRU
P4: ALU, BRU
P5: ALU, BRU
P6 (512 bits): store address, load
P7 (512 bits): store address, load
P8 (512 bits): store address, load
P9 (512 bits): store address, load
P10 (512 bits): VPU, DIV/SQRT
P11 (512 bits): VPU, DIV/SQRT
P12 (512 bits): VPU
P13 (512 bits): VPU
P14 (512 bits): FP store data
P15 (512 bits): FP store data

Cycles executing div or sqrt instructions

Cycles summary

Front-end	25.75
Dispatch	14.75
Overall L1	25.75

Vectorization ratios

INT

all	1%
load	0%
store	0%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	NA (no fma vectorizable/vectorized instructions)
other	3%

all	79%
load	78%
store	91%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	80%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	75%

INT+FP

all	46%
load	44%
store	50%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	57%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	43%

Vector efficiency ratios

INT

all	12%
load	12%
store	12%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	12%
fma	NA (no fma vectorizable/vectorized instructions)
other	12%

all	40%
load	37%
store	67%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	51%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	33%

INT+FP

all	28%
load	26%
store	42%
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	40%
fma	NA (no fma vectorizable/vectorized instructions)
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	23%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 25.75 cycles. At this rate:

7% of peak load performance is reached (19.73 out of 256.00 bytes loaded per cycle (GB/s @ 1GHz))
18% of peak store performance is reached (23.15 out of 128.00 bytes stored per cycle (GB/s @ 1GHz))

Front-end bottlenecks

ASM code

In the binary file, the address of the loop is: 6d7f

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	P15	Latency	Recip. throughput	Vectorization
MOV 0x380(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x2a8(%RSP),%R8	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x330(%RSP),%RDI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x220(%RSP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x238(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
INC %RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD %R8,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD %R8,%RSI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD %R8,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVSS %XMM2,(%R13,%RAX,4)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	1	0.50	scal (6.3%)
MOV %R14,%R8	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %RDX,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
CMP 0x38(%RSP),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
MOV %RCX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
JE 6ce0 <main+0x2290>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%R9	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %R14,%R10	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
AND $-0x8,%R9	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
AND $-0x40,%R10	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RAX,0x220(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6db0 <main+0x2360>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSS -0x5ed3(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (6.3%)
XOR %EAX,%EAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 6ec0 <main+0x2470>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x40,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6dd0 <main+0x2380>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVSS -0x5eee(%RIP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.50	scal (6.3%)
XOR %EAX,%EAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 6e6d <main+0x241d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
VBROADCASTSS -0x5f0a(%RIP),%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (6.3%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %ECX,%ECX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
ADD $-0x38,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
AND %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPS %ZMM0,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
VMOVAPS %ZMM0,%ZMM3	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (100.0%)
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
VMAXPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VMAXPS %ZMM3,%ZMM2,%ZMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VMAXPS %ZMM2,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF64X4 $0x1,%ZMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	4	0.50	vect (50.0%)
VMAXPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VMAXSS %XMM1,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
CMP %RAX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 6ecd <main+0x247d>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x38,%R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 6ec0 <main+0x2470>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV $0x7ffffffffffffff8,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %RAX,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND %RDX,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VMAXPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VMAXSS %XMM1,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
JMP 6ec8 <main+0x2478>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV %R14,0x380(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %RSI,0x28(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %R8,0x328(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %XMM1,0x2c0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
MOV %RDI,0x330(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV %R9,0x338(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6f10 <main+0x24c0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VXORPS %XMM2,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 7c20 <main+0x31d0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x40,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 6f30 <main+0x24e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VXORPS %XMM2,%XMM2,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
JMP 7a5a <main+0x300a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
VXORPS %XMM1,%XMM1,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R10,0x3e8(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
ADD $-0x38,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
AND %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,0x480(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VXORPS %XMM0,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
MOV %RAX,0x2b0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %ZMM0,0x340(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM0,0x400(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM0,0x440(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS 0x400(%RSP),%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (100.0%)
MOV 0x2b0(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VADDPS 0x340(%RSP),%ZMM0,%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPS 0x440(%RSP),%ZMM0,%ZMM0	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VADDPS %ZMM0,%ZMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF64X4 $0x1,%ZMM0,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	4	0.50	vect (50.0%)
VADDPS %ZMM1,%ZMM0,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (100.0%)
VEXTRACTF128 $0x1,%YMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VADDSS %XMM1,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
CMP %R13,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JNE 7a42 <main+0x2ff2>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
JMP 7c5e <main+0x320e>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
TEST $0x38,%R14B	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JE 7c20 <main+0x31d0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %R13,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV 0x338(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VXORPS %XMM0,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	vect (25.0%)
VBLENDPS $0x1,%XMM2,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
MOV %RAX,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
AND %RAX,%RCX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
MOV %RCX,0x2b0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
VMOVAPS %YMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
NOPL (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VEXTRACTF128 $0x1,%YMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%R14	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
MOV 0x2b0(%RSP),%R13	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
VADDPS %XMM0,%XMM2,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VSHUFPD $0x1,%XMM0,%XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (25.0%)
VADDPS %XMM1,%XMM0,%XMM0	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	vect (25.0%)
VMOVSHDUP %XMM0,%XMM1	1	0	0	0	0	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	1	0.25	vect (12.5%)
VADDSS %XMM1,%XMM0,%XMM2	1	0	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	2	0.50	scal (6.3%)
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
CMP %R13,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 7c5e <main+0x320e>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
NOPL (%RAX)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS %XMM2,0x240(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (25.0%)
CMP $0x8,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 7c80 <main+0x3230>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV 0x220(%RSP),%RCX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
XOR %R14D,%R14D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
JMP 8160 <main+0x3710>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
CMP $0x10,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JAE 7e40 <main+0x33f0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
XOR %R14D,%R14D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
VBROADCASTSS %XMM1,%YMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %R14,%R13	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %RAX,%R14	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM2,%YMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
AND %RAX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
VMOVAPS %YMM0,0xc0(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
VMOVAPS %YMM1,0x340(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0.50	0.50	4	0.50	vect (50.0%)
NOP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.09	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
CMP %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JNE 8160 <main+0x3710>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
JMP 81aa <main+0x375a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A
MOV %R14,%RAX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
AND $-0x10,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VBROADCASTSS %XMM1,%ZMM0	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
VBROADCASTSS %XMM2,%ZMM1	1	0	0	0	0	0	0	0	0	0	0	0	0.50	0.50	0	0	0	1	0.50	scal (6.3%)
MOV %R14,%RCX	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (12.5%)
XOR %R13D,%R13D	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
MOV %RAX,0x440(%RSP)	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	4	0.50	scal (12.5%)
MOV $0x7ffffffffffffff8,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
LEA -0x8(%RAX),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
VMOVAPS %ZMM0,0x340(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
VMOVAPS %ZMM1,0x400(%RSP)	2	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	1	1	4	1	vect (100.0%)
AND %RCX,%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
NOPW %CS:(%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
VMOVAPS 0x2c0(%RSP),%XMM1	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x380(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
CMP %R14,%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
JE 81aa <main+0x375a>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
TEST $0x8,%AL	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
JNE 7c90 <main+0x3240>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
NOPW (%RAX,%RAX,1)	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.13	N/A
MOV 0x328(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
LEA 0x1(%RAX),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
CMP 0x38(%RSP),%R14	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	scal (12.5%)
JAE 6d30 <main+0x22e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.33-1	N/A
MOV %RAX,%RDI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	N/A
IMUL 0x3e0(%RSP),%RDI	1	0.33	0.33	0.33	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.33	N/A
MOV 0x3d8(%RSP),%RDX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
LEA (,%RAX,4),%RAX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
XOR %ESI,%ESI	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.17	scal (6.3%)
SUB %RAX,%RDX	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	scal (12.5%)
ADD $0x4,%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0	0	0	0	0	0	0	0	0	0	1	0.17	N/A
ADD 0x80(%RSP),%RDI	1	0.17	0.17	0.17	0.17	0.17	0.17	0.25	0.25	0.25	0.25	0	0	0	0	0	0	1	0.25	N/A
VZEROUPPER	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	vect (25.0%)
CALL 89c0 <@plt_start@+0x30>	2	0	0	0	0.33	0.33	0.33	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0.50	N/A
VMOVAPS 0x240(%RSP),%XMM2	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.50	vect (25.0%)
MOV 0x328(%RSP),%RAX	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	N/A
MOV 0x28(%RSP),%RSI	1	0	0	0	0	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	3	0.25	scal (12.5%)
JMP 6d30 <main+0x22e0>	1	0	0	0	0.33	0.33	0.33	0	0	0	0	0	0	0	0	0	0	0	0.50	N/A

Report Configuration

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

CALL instructions

Details

Complex instructions

Details

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code

Code clean check

Workaround

Vectorization

Details

Workaround

Execution units bottlenecks

CALL instructions

Details

Complex instructions

Details

Vector unaligned load/store instructions

Details

Workaround

Type of elements and instruction set

Matching between your loop (in the source code) and the binary loop

Arithmetic intensity

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Vector efficiency ratios

Cycles and memory resources usage

Front-end bottlenecks

ASM code