DIVPS (XMM, XMM)

Experiment 1

Instruction: DIVPS XMM3, XMM4
Chain instructions: ORPS XMM3, XMM1; ANDPS XMM3, XMM1; ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3
Chain latency: ≥10

Code:

   0:	0f 5e dc             	divps  xmm3,xmm4
   3:	0f 56 d9             	orps   xmm3,xmm1
   6:	0f 54 d9             	andps  xmm3,xmm1
   9:	0f 56 db             	orps   xmm3,xmm3
   c:	0f 56 db             	orps   xmm3,xmm3
   f:	0f 56 db             	orps   xmm3,xmm3
  12:	0f 56 db             	orps   xmm3,xmm3
  15:	0f 56 db             	orps   xmm3,xmm3
  18:	0f 56 db             	orps   xmm3,xmm3
  1b:	0f 56 db             	orps   xmm3,xmm3
  1e:	0f 56 db             	orps   xmm3,xmm3

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 11.0
- Core cycles: 22.0
- Reference cycles: 23.07
- UOPS_EXECUTED.THREAD: 11.0

Experiment 2

Instruction: DIVPS XMM3, XMM3

Code:

   0:	0f 5e db             	divps  xmm3,xmm3

Init:

VZEROALL;
MOV RAX, 0x3f8000003f800000;
MOV RBX, 0x3f8000003f800000;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 1.0
- Core cycles: 11.0
- Reference cycles: 11.6
- UOPS_EXECUTED.THREAD: 1.0

Experiment 3

Instruction: DIVPS XMM3, XMM4
Chain instructions: ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3;ORPS XMM3, XMM3
Chain latency: ≥8

Code:

   0:	0f 5e dc             	divps  xmm3,xmm4
   3:	0f 56 db             	orps   xmm3,xmm3
   6:	0f 56 db             	orps   xmm3,xmm3
   9:	0f 56 db             	orps   xmm3,xmm3
   c:	0f 56 db             	orps   xmm3,xmm3
   f:	0f 56 db             	orps   xmm3,xmm3
  12:	0f 56 db             	orps   xmm3,xmm3
  15:	0f 56 db             	orps   xmm3,xmm3
  18:	0f 56 db             	orps   xmm3,xmm3

Init:

VZEROALL;
MOV RAX, 0x3f8000003f800000;
MOV RBX, 0x3f8000003f800000;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 9.0
- Core cycles: 19.25
- Reference cycles: 20.4
- UOPS_EXECUTED.THREAD: 9.0

Experiment 1

Instruction: DIVPS XMM3, XMM4
Chain instructions: ANDPS XMM4, XMM3; MOVUPS XMM3, XMM1; ANDPS XMM4, XMM2; ORPS XMM4, XMM2; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4;
Chain latency: ≥11

Code:

   0:	0f 5e dc             	divps  xmm3,xmm4
   3:	0f 54 e3             	andps  xmm4,xmm3
   6:	0f 10 d9             	movups xmm3,xmm1
   9:	0f 54 e2             	andps  xmm4,xmm2
   c:	0f 56 e2             	orps   xmm4,xmm2
   f:	0f 56 e4             	orps   xmm4,xmm4
  12:	0f 56 e4             	orps   xmm4,xmm4
  15:	0f 56 e4             	orps   xmm4,xmm4
  18:	0f 56 e4             	orps   xmm4,xmm4
  1b:	0f 56 e4             	orps   xmm4,xmm4
  1e:	0f 56 e4             	orps   xmm4,xmm4
  21:	0f 56 e4             	orps   xmm4,xmm4
  24:	0f 56 e4             	orps   xmm4,xmm4

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 13.0
- Core cycles: 22.0
- Reference cycles: 23.2
- UOPS_EXECUTED.THREAD: 12.0

Experiment 2

Instruction: DIVPS XMM3, XMM3

Code:

   0:	0f 5e db             	divps  xmm3,xmm3

Init:

VZEROALL;
MOV RAX, 0x3f8000003f800000;
MOV RBX, 0x3f8000003f800000;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 1.0
- Core cycles: 11.0
- Reference cycles: 11.6
- UOPS_EXECUTED.THREAD: 1.0

Experiment 3

Instruction: DIVPS XMM3, XMM4
Chain instructions: ANDPS XMM4, XMM3; ANDPS XMM4, XMM2; ORPS XMM4, XMM2; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4; ORPS XMM4, XMM4;
Chain latency: ≥11

Code:

   0:	0f 5e dc             	divps  xmm3,xmm4
   3:	0f 54 e3             	andps  xmm4,xmm3
   6:	0f 54 e2             	andps  xmm4,xmm2
   9:	0f 56 e2             	orps   xmm4,xmm2
   c:	0f 56 e4             	orps   xmm4,xmm4
   f:	0f 56 e4             	orps   xmm4,xmm4
  12:	0f 56 e4             	orps   xmm4,xmm4
  15:	0f 56 e4             	orps   xmm4,xmm4
  18:	0f 56 e4             	orps   xmm4,xmm4
  1b:	0f 56 e4             	orps   xmm4,xmm4
  1e:	0f 56 e4             	orps   xmm4,xmm4
  21:	0f 56 e4             	orps   xmm4,xmm4

Init:

VZEROALL;
MOV RAX, 0x3f8000003f800000;
MOV RBX, 0x3f8000003f800000;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, [R14];
MOVUPS XMM3, [R14+64];
MOVUPS XMM4, [R14]

Show nanoBench command
Results:
- Instructions retired: 12.0
- Core cycles: 22.0
- Reference cycles: 23.2
- UOPS_EXECUTED.THREAD: 12.0

DIVPS (XMM, XMM) - Latency

Operands

Latency operand 1 → 1: ≤11

Latency operand 2 → 1: ≤11

Latency operand 1 → 1: ≤11

Experiment 1

Experiment 2

Experiment 3

Latency operand 2 → 1: ≤11

Experiment 1

Experiment 2

Experiment 3