PCLMULQDQ (XMM, XMM, I8) - Port Usage

With blocking instructions for port '1':

Code:

   0:	66 41 0f 3a 20 d0 02 	pinsrb xmm2,r8d,0x2
   7:	66 41 0f 3a 20 d8 02 	pinsrb xmm3,r8d,0x2
   e:	66 41 0f 3a 20 e0 02 	pinsrb xmm4,r8d,0x2
  15:	66 41 0f 3a 20 e8 02 	pinsrb xmm5,r8d,0x2
  1c:	66 41 0f 3a 20 f0 02 	pinsrb xmm6,r8d,0x2
  23:	66 41 0f 3a 20 f8 02 	pinsrb xmm7,r8d,0x2
  2a:	66 45 0f 3a 20 c0 02 	pinsrb xmm8,r8d,0x2
  31:	66 45 0f 3a 20 c8 02 	pinsrb xmm9,r8d,0x2
  38:	66 45 0f 3a 20 d0 02 	pinsrb xmm10,r8d,0x2
  3f:	66 45 0f 3a 20 d8 02 	pinsrb xmm11,r8d,0x2
  46:	66 0f 3a 44 c1 02    	pclmullqhqdq xmm0,xmm1

Init:

VZEROALL;
PXOR XMM2, XMM2;
PXOR XMM3, XMM3;
PXOR XMM4, XMM4;
PXOR XMM5, XMM5;
PXOR XMM6, XMM6;
PXOR XMM7, XMM7;
PXOR XMM8, XMM8;
PXOR XMM9, XMM9;
PXOR XMM10, XMM10;
PXOR XMM11, XMM11;
PXOR XMM12, XMM12;
PXOR XMM0, XMM0;
PXOR XMM1, XMM1

Show nanoBench command
Results:
- MPERF: 9.19
- APERF: 11.08
- FpuPipeAssignment.Total0: 1.0
- FpuPipeAssignment.Total1: 10.99
- FpuPipeAssignment.Total2: 1.01
- FpuPipeAssignment.Total3: 0.0
- FpuPipeAssignment.Total4: 0.0
- FpuPipeAssignment.Total5: 0.0

⇨ One μop that can only use port '1'

With blocking instructions for ports {'0', '1'}:

Code:

   0:	66 0f 38 de d3       	aesdec xmm2,xmm3
   5:	66 0f 38 de e3       	aesdec xmm4,xmm3
   a:	66 0f 38 de eb       	aesdec xmm5,xmm3
   f:	66 0f 38 de f3       	aesdec xmm6,xmm3
  14:	66 0f 38 de fb       	aesdec xmm7,xmm3
  19:	66 44 0f 38 de c3    	aesdec xmm8,xmm3
  1f:	66 44 0f 38 de cb    	aesdec xmm9,xmm3
  25:	66 44 0f 38 de d3    	aesdec xmm10,xmm3
  2b:	66 44 0f 38 de db    	aesdec xmm11,xmm3
  31:	66 44 0f 38 de e3    	aesdec xmm12,xmm3
  37:	66 0f 38 de d3       	aesdec xmm2,xmm3
  3c:	66 0f 38 de e3       	aesdec xmm4,xmm3
  41:	66 0f 38 de eb       	aesdec xmm5,xmm3
  46:	66 0f 38 de f3       	aesdec xmm6,xmm3
  4b:	66 0f 38 de fb       	aesdec xmm7,xmm3
  50:	66 44 0f 38 de c3    	aesdec xmm8,xmm3
  56:	66 0f 3a 44 c1 02    	pclmullqhqdq xmm0,xmm1

Init:

VZEROALL;
PXOR XMM2, XMM2;
PXOR XMM3, XMM3;
PXOR XMM4, XMM4;
PXOR XMM5, XMM5;
PXOR XMM6, XMM6;
PXOR XMM7, XMM7;
PXOR XMM8, XMM8;
PXOR XMM9, XMM9;
PXOR XMM10, XMM10;
PXOR XMM11, XMM11;
PXOR XMM12, XMM12;
PXOR XMM0, XMM0;
PXOR XMM1, XMM1

Show nanoBench command
Results:
- MPERF: 7.03
- APERF: 8.63
- FpuPipeAssignment.Total0: 8.5
- FpuPipeAssignment.Total1: 8.53
- FpuPipeAssignment.Total2: 1.97
- FpuPipeAssignment.Total3: 0.0
- FpuPipeAssignment.Total4: 0.0
- FpuPipeAssignment.Total5: 0.0

With blocking instructions for ports {'1', '2'}:

Code:

   0:	66 0f 3a 21 d3 02    	insertps xmm2,xmm3,0x2
   6:	66 0f 3a 21 e3 02    	insertps xmm4,xmm3,0x2
   c:	66 0f 3a 21 eb 02    	insertps xmm5,xmm3,0x2
  12:	66 0f 3a 21 f3 02    	insertps xmm6,xmm3,0x2
  18:	66 0f 3a 21 fb 02    	insertps xmm7,xmm3,0x2
  1e:	66 44 0f 3a 21 c3 02 	insertps xmm8,xmm3,0x2
  25:	66 44 0f 3a 21 cb 02 	insertps xmm9,xmm3,0x2
  2c:	66 44 0f 3a 21 d3 02 	insertps xmm10,xmm3,0x2
  33:	66 44 0f 3a 21 db 02 	insertps xmm11,xmm3,0x2
  3a:	66 44 0f 3a 21 e3 02 	insertps xmm12,xmm3,0x2
  41:	66 0f 3a 21 d3 02    	insertps xmm2,xmm3,0x2
  47:	66 0f 3a 21 e3 02    	insertps xmm4,xmm3,0x2
  4d:	66 0f 3a 21 eb 02    	insertps xmm5,xmm3,0x2
  53:	66 0f 3a 21 f3 02    	insertps xmm6,xmm3,0x2
  59:	66 0f 3a 21 fb 02    	insertps xmm7,xmm3,0x2
  5f:	66 44 0f 3a 21 c3 02 	insertps xmm8,xmm3,0x2
  66:	66 0f 3a 44 c1 02    	pclmullqhqdq xmm0,xmm1

Init:

VZEROALL;
MOV RAX, 0x4000000040000000;
MOV [R14+0], RAX;
MOV [R14+8], RAX;
MOVUPD XMM2, [R14];
MOVUPD XMM3, [R14];
MOVUPD XMM4, [R14];
MOVUPD XMM5, [R14];
MOVUPD XMM6, [R14];
MOVUPD XMM7, [R14];
MOVUPD XMM8, [R14];
MOVUPD XMM9, [R14];
MOVUPD XMM10, [R14];
MOVUPD XMM11, [R14];
MOVUPD XMM12, [R14];
PXOR XMM0, XMM0;
PXOR XMM1, XMM1

Show nanoBench command
Results:
- MPERF: 7.4
- APERF: 9.07
- FpuPipeAssignment.Total0: 1.0
- FpuPipeAssignment.Total1: 9.01
- FpuPipeAssignment.Total2: 8.99
- FpuPipeAssignment.Total3: 0.0
- FpuPipeAssignment.Total4: 0.0
- FpuPipeAssignment.Total5: 0.0

⇨ One μop that can only use ports {'1', '2'}

With blocking instructions for ports {'0', '1', '2', '3'}:

Code:

   0:	66 0f 55 d3          	andnpd xmm2,xmm3
   4:	66 0f 55 e3          	andnpd xmm4,xmm3
   8:	66 0f 55 eb          	andnpd xmm5,xmm3
   c:	66 0f 55 f3          	andnpd xmm6,xmm3
  10:	66 0f 55 fb          	andnpd xmm7,xmm3
  14:	66 44 0f 55 c3       	andnpd xmm8,xmm3
  19:	66 44 0f 55 cb       	andnpd xmm9,xmm3
  1e:	66 44 0f 55 d3       	andnpd xmm10,xmm3
  23:	66 44 0f 55 db       	andnpd xmm11,xmm3
  28:	66 44 0f 55 e3       	andnpd xmm12,xmm3
  2d:	66 0f 55 d3          	andnpd xmm2,xmm3
  31:	66 0f 55 e3          	andnpd xmm4,xmm3
  35:	66 0f 55 eb          	andnpd xmm5,xmm3
  39:	66 0f 55 f3          	andnpd xmm6,xmm3
  3d:	66 0f 55 fb          	andnpd xmm7,xmm3
  41:	66 44 0f 55 c3       	andnpd xmm8,xmm3
  46:	66 44 0f 55 cb       	andnpd xmm9,xmm3
  4b:	66 44 0f 55 d3       	andnpd xmm10,xmm3
  50:	66 44 0f 55 db       	andnpd xmm11,xmm3
  55:	66 44 0f 55 e3       	andnpd xmm12,xmm3
  5a:	66 0f 55 d3          	andnpd xmm2,xmm3
  5e:	66 0f 55 e3          	andnpd xmm4,xmm3
  62:	66 0f 55 eb          	andnpd xmm5,xmm3
  66:	66 0f 55 f3          	andnpd xmm6,xmm3
  6a:	66 0f 55 fb          	andnpd xmm7,xmm3
  6e:	66 44 0f 55 c3       	andnpd xmm8,xmm3
  73:	66 44 0f 55 cb       	andnpd xmm9,xmm3
  78:	66 44 0f 55 d3       	andnpd xmm10,xmm3
  7d:	66 44 0f 55 db       	andnpd xmm11,xmm3
  82:	66 44 0f 55 e3       	andnpd xmm12,xmm3
  87:	66 0f 55 d3          	andnpd xmm2,xmm3
  8b:	66 0f 55 e3          	andnpd xmm4,xmm3
  8f:	66 0f 3a 44 c1 02    	pclmullqhqdq xmm0,xmm1

Init:

VZEROALL;
PXOR XMM2, XMM2;
PXOR XMM3, XMM3;
PXOR XMM4, XMM4;
PXOR XMM5, XMM5;
PXOR XMM6, XMM6;
PXOR XMM7, XMM7;
PXOR XMM8, XMM8;
PXOR XMM9, XMM9;
PXOR XMM10, XMM10;
PXOR XMM11, XMM11;
PXOR XMM12, XMM12;
PXOR XMM0, XMM0;
PXOR XMM1, XMM1

Show nanoBench command
Results:
- MPERF: 8.14
- APERF: 9.79
- FpuPipeAssignment.Total0: 8.85
- FpuPipeAssignment.Total1: 9.01
- FpuPipeAssignment.Total2: 9.0
- FpuPipeAssignment.Total3: 8.14
- FpuPipeAssignment.Total4: 0.0
- FpuPipeAssignment.Total5: 0.0

⇨ One μop that can only use ports {'0', '1', '2', '3'}