DIVPS (XMM, M128) - Port Usage

With lfence (to avoid incorrect counts due to replays)

Code:

   0:	0f 10 d1             	movups xmm2,xmm1
   3:	41 0f 5e 16          	divps  xmm2,XMMWORD PTR [r14]
   7:	90                   	nop
   8:	90                   	nop
   9:	90                   	nop
   a:	90                   	nop
   b:	90                   	nop
   c:	90                   	nop
   d:	90                   	nop
   e:	90                   	nop
   f:	90                   	nop
  10:	90                   	nop
  11:	0f ae e8             	lfence

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 13.0
- Core cycles: 26.0
- Reference cycles: 18.52
- UOPS_DISPATCHED.INT_EU_ALL: 1.01
- UOPS_DISPATCHED.ALU: 1.0
- UOPS_DISPATCHED.LD: 1.0
- UOPS_DISPATCHED.SLOW: 0.0
- UOPS_DISPATCHED.STD: 0.0
- UOPS_DISPATCHED.SHIFT: 0.0
- UOPS_DISPATCHED.JMP: 0.0
- UOPS_DISPATCHED.STA: 0.0

Code:

   0:	0f 10 d1             	movups xmm2,xmm1

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 1.0
- Core cycles: 0.12
- Reference cycles: -0.16
- UOPS_DISPATCHED.INT_EU_ALL: 0.0
- UOPS_DISPATCHED.ALU: 0.0
- UOPS_DISPATCHED.LD: 0.0
- UOPS_DISPATCHED.SLOW: 0.0
- UOPS_DISPATCHED.STD: 0.0
- UOPS_DISPATCHED.SHIFT: 0.0
- UOPS_DISPATCHED.JMP: 0.0
- UOPS_DISPATCHED.STA: 0.0

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

Code:

   0:	66 41 0f 5f 46 40    	maxpd  xmm0,XMMWORD PTR [r14+0x40]
   6:	66 41 0f 5f 4e 50    	maxpd  xmm1,XMMWORD PTR [r14+0x50]
   c:	66 41 0f 5f 5e 60    	maxpd  xmm3,XMMWORD PTR [r14+0x60]
  12:	66 41 0f 5f 66 70    	maxpd  xmm4,XMMWORD PTR [r14+0x70]
  18:	66 41 0f 5f ae 80 00 	maxpd  xmm5,XMMWORD PTR [r14+0x80]
  1f:	00 00 
  21:	66 41 0f 5f b6 90 00 	maxpd  xmm6,XMMWORD PTR [r14+0x90]
  28:	00 00 
  2a:	66 41 0f 5f be a0 00 	maxpd  xmm7,XMMWORD PTR [r14+0xa0]
  31:	00 00 
  33:	66 45 0f 5f 86 b0 00 	maxpd  xmm8,XMMWORD PTR [r14+0xb0]
  3a:	00 00 
  3c:	66 45 0f 5f 8e c0 00 	maxpd  xmm9,XMMWORD PTR [r14+0xc0]
  43:	00 00 
  45:	66 45 0f 5f 96 d0 00 	maxpd  xmm10,XMMWORD PTR [r14+0xd0]
  4c:	00 00 
  4e:	0f 10 d1             	movups xmm2,xmm1
  51:	41 0f 5e 16          	divps  xmm2,XMMWORD PTR [r14]

Init:

VZEROALL;
MOV RAX, 0x4000000040000000;
MOV [R14+0], RAX;
MOV [R14+8], RAX;
MOVUPD XMM0, [R14];
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPD XMM1, [R14];
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOVUPD XMM3, [R14];
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOVUPD XMM4, [R14];
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPD XMM5, [R14];
MOV [R14+128], RAX;
MOV [R14+136], RAX;
MOVUPD XMM6, [R14];
MOV [R14+144], RAX;
MOV [R14+152], RAX;
MOVUPD XMM7, [R14];
MOV [R14+160], RAX;
MOV [R14+168], RAX;
MOVUPD XMM8, [R14];
MOV [R14+176], RAX;
MOV [R14+184], RAX;
MOVUPD XMM9, [R14];
MOV [R14+192], RAX;
MOV [R14+200], RAX;
MOVUPD XMM10, [R14];
MOV [R14+208], RAX;
MOV [R14+216], RAX;
MOVUPD XMM11, [R14];
MOV [R14+224], RAX;
MOV [R14+232], RAX;
MOVUPD XMM12, [R14];
MOV [R14+240], RAX;
MOV [R14+248], RAX;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 12.0
- Core cycles: 6.35
- Reference cycles: 4.38
- UOPS_DISPATCHED.V0: 5.47
- UOPS_DISPATCHED.V1: 5.53
- UOPS_DISPATCHED.V2: 0.0
- UOPS_DISPATCHED.V3: 0.0

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

With an indexed addressing mode

With lfence (to avoid incorrect counts due to replays)

Code:

   0:	0f 10 d1             	movups xmm2,xmm1
   3:	41 0f 5e 16          	divps  xmm2,XMMWORD PTR [r14]
   7:	90                   	nop
   8:	90                   	nop
   9:	90                   	nop
   a:	90                   	nop
   b:	90                   	nop
   c:	90                   	nop
   d:	90                   	nop
   e:	90                   	nop
   f:	90                   	nop
  10:	90                   	nop
  11:	0f ae e8             	lfence

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 13.0
- Core cycles: 26.0
- Reference cycles: 18.01
- UOPS_DISPATCHED.INT_EU_ALL: 1.01
- UOPS_DISPATCHED.ALU: 1.01
- UOPS_DISPATCHED.LD: 1.0
- UOPS_DISPATCHED.SLOW: 0.0
- UOPS_DISPATCHED.STD: 0.0
- UOPS_DISPATCHED.SHIFT: 0.0
- UOPS_DISPATCHED.JMP: 0.0
- UOPS_DISPATCHED.STA: 0.0

Code:

   0:	0f 10 d1             	movups xmm2,xmm1

Init:

VZEROALL;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 1.0
- Core cycles: 0.12
- Reference cycles: 0.34
- UOPS_DISPATCHED.INT_EU_ALL: 0.0
- UOPS_DISPATCHED.ALU: 0.0
- UOPS_DISPATCHED.LD: 0.0
- UOPS_DISPATCHED.SLOW: 0.0
- UOPS_DISPATCHED.STD: 0.0
- UOPS_DISPATCHED.SHIFT: 0.0
- UOPS_DISPATCHED.JMP: 0.0
- UOPS_DISPATCHED.STA: 0.0

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

Code:

   0:	66 41 0f 5f 46 40    	maxpd  xmm0,XMMWORD PTR [r14+0x40]
   6:	66 41 0f 5f 4e 50    	maxpd  xmm1,XMMWORD PTR [r14+0x50]
   c:	66 41 0f 5f 5e 60    	maxpd  xmm3,XMMWORD PTR [r14+0x60]
  12:	66 41 0f 5f 66 70    	maxpd  xmm4,XMMWORD PTR [r14+0x70]
  18:	66 41 0f 5f ae 80 00 	maxpd  xmm5,XMMWORD PTR [r14+0x80]
  1f:	00 00 
  21:	66 41 0f 5f b6 90 00 	maxpd  xmm6,XMMWORD PTR [r14+0x90]
  28:	00 00 
  2a:	66 41 0f 5f be a0 00 	maxpd  xmm7,XMMWORD PTR [r14+0xa0]
  31:	00 00 
  33:	66 45 0f 5f 86 b0 00 	maxpd  xmm8,XMMWORD PTR [r14+0xb0]
  3a:	00 00 
  3c:	66 45 0f 5f 8e c0 00 	maxpd  xmm9,XMMWORD PTR [r14+0xc0]
  43:	00 00 
  45:	66 45 0f 5f 96 d0 00 	maxpd  xmm10,XMMWORD PTR [r14+0xd0]
  4c:	00 00 
  4e:	0f 10 d1             	movups xmm2,xmm1
  51:	41 0f 5e 16          	divps  xmm2,XMMWORD PTR [r14]

Init:

VZEROALL;
MOV RAX, 0x4000000040000000;
MOV [R14+0], RAX;
MOV [R14+8], RAX;
MOVUPD XMM0, [R14];
MOV [R14+64], RAX;
MOV [R14+72], RAX;
MOVUPD XMM1, [R14];
MOV [R14+80], RAX;
MOV [R14+88], RAX;
MOVUPD XMM3, [R14];
MOV [R14+96], RAX;
MOV [R14+104], RAX;
MOVUPD XMM4, [R14];
MOV [R14+112], RAX;
MOV [R14+120], RAX;
MOVUPD XMM5, [R14];
MOV [R14+128], RAX;
MOV [R14+136], RAX;
MOVUPD XMM6, [R14];
MOV [R14+144], RAX;
MOV [R14+152], RAX;
MOVUPD XMM7, [R14];
MOV [R14+160], RAX;
MOV [R14+168], RAX;
MOVUPD XMM8, [R14];
MOV [R14+176], RAX;
MOV [R14+184], RAX;
MOVUPD XMM9, [R14];
MOV [R14+192], RAX;
MOV [R14+200], RAX;
MOVUPD XMM10, [R14];
MOV [R14+208], RAX;
MOV [R14+216], RAX;
MOVUPD XMM11, [R14];
MOV [R14+224], RAX;
MOV [R14+232], RAX;
MOVUPD XMM12, [R14];
MOV [R14+240], RAX;
MOV [R14+248], RAX;
MOV RAX, 0x54ed392654ed3926;
MOV RBX, 0x3f99f4c03f99f4c0;
MOV [R14+0], RBX;
MOV [R14+8], RBX;
MOV [R14+16], RBX;
MOV [R14+24], RBX;
MOV [R14+32], RBX;
MOV [R14+40], RBX;
MOV [R14+48], RBX;
MOV [R14+56], RBX;
MOVUPS XMM0, [R14];
MOVUPS XMM1, [R14+64];
MOVUPS XMM2, XMM1

Show nanoBench command
Results:
- Instructions retired: 12.0
- Core cycles: 6.32
- Reference cycles: 4.38
- UOPS_DISPATCHED.V0: 5.51
- UOPS_DISPATCHED.V1: 5.5
- UOPS_DISPATCHED.V2: 0.0
- UOPS_DISPATCHED.V3: 0.0

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