MOVUPS (XMM, M128) - Latency

Experiment 1

• Instruction: MOVUPS XMM0, xmmword ptr [R14]
• Chain instructions: MOVQ R12, XMM0;XOR R14, R12;XOR R14, R12;XOR R14, R12;XOR R14, R12
• Chain latency: ≥5
• Code:

     0:	41 0f 10 06          	movups xmm0,XMMWORD PTR [r14]
     4:	66 49 0f 7e c4       	movq   r12,xmm0
     9:	4d 31 e6             	xor    r14,r12
     c:	4d 31 e6             	xor    r14,r12
     f:	4d 31 e6             	xor    r14,r12
    12:	4d 31 e6             	xor    r14,r12
  

• Init:

  MOV RAX, 0x4000000040000000;
  MOV [R14+0], RAX;
  MOV [R14+8], RAX

• Show nanoBench command
• Results:
  • Instructions retired: 6.0
  • Core cycles: 12.0
  • Reference cycles: 7.8
  • UOPS_RETIRED.ANY: 6.0

Experiment 1

• Instruction: MOVUPS XMM0, xmmword ptr [R14+R13]
• Chain instructions: MOVQ R12, XMM0;XOR R13, R12;XOR R13, R12;XOR R13, R12;XOR R13, R12
• Chain latency: ≥5
• Code:

     0:	43 0f 10 04 2e       	movups xmm0,XMMWORD PTR [r14+r13*1]
     5:	66 49 0f 7e c4       	movq   r12,xmm0
     a:	4d 31 e5             	xor    r13,r12
     d:	4d 31 e5             	xor    r13,r12
    10:	4d 31 e5             	xor    r13,r12
    13:	4d 31 e5             	xor    r13,r12
  

• Init:

  MOV RAX, 0x4000000040000000;
  MOV [R14+0], RAX;
  MOV [R14+8], RAX;
  XOR R13, R13

• Show nanoBench command
• Results:
  • Instructions retired: 6.0
  • Core cycles: 12.0
  • Reference cycles: 7.8
  • UOPS_RETIRED.ANY: 6.0

Experiment 1

• Instruction: MOVUPS XMM0, xmmword ptr [R14]
• Chain instructions: SHUFPD XMM13, XMM0, 0;SHUFPD XMM13, XMM13, 0;SHUFPD XMM13, XMM13, 0;SHUFPD XMM13, XMM13, 0;SHUFPD XMM13, XMM13, 0;MOVUPD [R14], XMM13
• Chain latency: ≥6
• Code:

     0:	41 0f 10 06          	movups xmm0,XMMWORD PTR [r14]
     4:	66 44 0f c6 e8 00    	shufpd xmm13,xmm0,0x0
     a:	66 45 0f c6 ed 00    	shufpd xmm13,xmm13,0x0
    10:	66 45 0f c6 ed 00    	shufpd xmm13,xmm13,0x0
    16:	66 45 0f c6 ed 00    	shufpd xmm13,xmm13,0x0
    1c:	66 45 0f c6 ed 00    	shufpd xmm13,xmm13,0x0
    22:	66 45 0f 11 2e       	movupd XMMWORD PTR [r14],xmm13
  

• Init:

  MOV RAX, 0x4000000040000000;
  MOV [R14+0], RAX;
  MOV [R14+8], RAX

• Show nanoBench command
• Results:
  • Instructions retired: 7.0
  • Core cycles: 10.99
  • Reference cycles: 7.02
  • UOPS_RETIRED.ANY: 7.0

Experiment 2

• Instruction: MOVUPS XMM0, xmmword ptr [R14]
• Chain instructions: PSHUFD XMM13, XMM0, 0;PAND XMM13, XMM13;PAND XMM13, XMM13;PAND XMM13, XMM13;PAND XMM13, XMM13;MOVDQU [R14], XMM13
• Chain latency: ≥6
• Code:

     0:	41 0f 10 06          	movups xmm0,XMMWORD PTR [r14]
     4:	66 44 0f 70 e8 00    	pshufd xmm13,xmm0,0x0
     a:	66 45 0f db ed       	pand   xmm13,xmm13
     f:	66 45 0f db ed       	pand   xmm13,xmm13
    14:	66 45 0f db ed       	pand   xmm13,xmm13
    19:	66 45 0f db ed       	pand   xmm13,xmm13
    1e:	f3 45 0f 7f 2e       	movdqu XMMWORD PTR [r14],xmm13
  

• Init:

  MOV RAX, 0x4000000040000000;
  MOV [R14+0], RAX;
  MOV [R14+8], RAX

• Show nanoBench command
• Results:
  • Instructions retired: 7.0
  • Core cycles: 11.74
  • Reference cycles: 7.54
  • UOPS_RETIRED.ANY: 7.0