VPERMQ (YMM, M256, I8) - Latency

Experiment 1

• Instruction: VPERMQ YMM0, ymmword ptr [R14], 2
• Chain instructions: VMOVQ R12, XMM0;XOR R14, R12;XOR R14, R12;XOR R14, R12;XOR R14, R12
• Chain latency: ≥5
• Code:

     0:	c4 c3 fd 00 06 02    	vpermq ymm0,YMMWORD PTR [r14],0x2
     6:	c4 c1 f9 7e c4       	vmovq  r12,xmm0
     b:	4d 31 e6             	xor    r14,r12
     e:	4d 31 e6             	xor    r14,r12
    11:	4d 31 e6             	xor    r14,r12
    14:	4d 31 e6             	xor    r14,r12
  

• Init:

  MOV R15, 10000;
  L: VADDPS YMM0, YMM1, YMM1;
  VADDPS YMM0, YMM1, YMM1;
  DEC R15;
  JNZ L;
  VZEROALL

• Show nanoBench command
• Results:
  • MPERF: 15.24
  • APERF: 18.0
  • UOPS: 8.0

Experiment 1

• Instruction: VPERMQ YMM0, ymmword ptr [R14+R13], 2
• Chain instructions: VMOVQ R12, XMM0;XOR R13, R12;XOR R13, R12;XOR R13, R12;XOR R13, R12
• Chain latency: ≥5
• Code:

     0:	c4 83 fd 00 04 2e 02 	vpermq ymm0,YMMWORD PTR [r14+r13*1],0x2
     7:	c4 c1 f9 7e c4       	vmovq  r12,xmm0
     c:	4d 31 e5             	xor    r13,r12
     f:	4d 31 e5             	xor    r13,r12
    12:	4d 31 e5             	xor    r13,r12
    15:	4d 31 e5             	xor    r13,r12
  

• Init:

  MOV R15, 10000;
  L: VADDPS YMM0, YMM1, YMM1;
  VADDPS YMM0, YMM1, YMM1;
  DEC R15;
  JNZ L;
  VZEROALL;
  XOR R13, R13

• Show nanoBench command
• Results:
  • MPERF: 15.12
  • APERF: 18.0
  • UOPS: 8.0

Experiment 1

• Instruction: VPERMQ YMM0, ymmword ptr [R14], 2
• Chain instructions: VANDPD YMM13, YMM0, YMM0;VANDPD YMM13, YMM13, YMM13;VANDPD YMM13, YMM13, YMM13;VANDPD YMM13, YMM13, YMM13;VANDPD YMM13, YMM13, YMM13;VMOVUPD [R14], YMM13
• Chain latency: ≥6
• Code:

     0:	c4 c3 fd 00 06 02    	vpermq ymm0,YMMWORD PTR [r14],0x2
     6:	c5 7d 54 e8          	vandpd ymm13,ymm0,ymm0
     a:	c4 41 15 54 ed       	vandpd ymm13,ymm13,ymm13
     f:	c4 41 15 54 ed       	vandpd ymm13,ymm13,ymm13
    14:	c4 41 15 54 ed       	vandpd ymm13,ymm13,ymm13
    19:	c4 41 15 54 ed       	vandpd ymm13,ymm13,ymm13
    1e:	c4 41 7d 11 2e       	vmovupd YMMWORD PTR [r14],ymm13
  

• Init:

  MOV R15, 10000;
  L: VADDPS YMM0, YMM1, YMM1;
  VADDPS YMM0, YMM1, YMM1;
  DEC R15;
  JNZ L;
  VZEROALL

• Show nanoBench command
• Results:
  • MPERF: 16.68
  • APERF: 20.0
  • UOPS: 9.0

Experiment 2

• Instruction: VPERMQ YMM0, ymmword ptr [R14], 2
• Chain instructions: VPAND YMM13, YMM0, YMM0;VPAND YMM13, YMM13, YMM13;VPAND YMM13, YMM13, YMM13;VPAND YMM13, YMM13, YMM13;VPAND YMM13, YMM13, YMM13;VMOVDQU [R14], YMM13
• Chain latency: ≥6
• Code:

     0:	c4 c3 fd 00 06 02    	vpermq ymm0,YMMWORD PTR [r14],0x2
     6:	c5 7d db e8          	vpand  ymm13,ymm0,ymm0
     a:	c4 41 15 db ed       	vpand  ymm13,ymm13,ymm13
     f:	c4 41 15 db ed       	vpand  ymm13,ymm13,ymm13
    14:	c4 41 15 db ed       	vpand  ymm13,ymm13,ymm13
    19:	c4 41 15 db ed       	vpand  ymm13,ymm13,ymm13
    1e:	c4 41 7e 7f 2e       	vmovdqu YMMWORD PTR [r14],ymm13
  

• Init:

  MOV R15, 10000;
  L: VADDPS YMM0, YMM1, YMM1;
  VADDPS YMM0, YMM1, YMM1;
  DEC R15;
  JNZ L;
  VZEROALL

• Show nanoBench command
• Results:
  • MPERF: 15.96
  • APERF: 19.0
  • UOPS: 9.0