VPSHUFBITQMB (K, K, YMM, YMM) - Latency

Experiment 1

• Instruction: VPSHUFBITQMB K2 {K1}, YMM0, YMM1
• Chain instruction: VPSHUFBITQMB K1 {K2}, YMM0, YMM1
• Code:

     0:	62 f2 7d 29 8f d1    	vpshufbitqmb k2{k1},ymm0,ymm1
     6:	62 f2 7d 2a 8f c9    	vpshufbitqmb k1{k2},ymm0,ymm1
  

• Init:

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

• Show nanoBench command
• Results:
  • MPERF: 5.38
  • APERF: 8.0
  • UOPS: 2.0

Experiment 2

• Instruction: VPSHUFBITQMB K2 {K1}, YMM0, YMM1
• Chain instructions: KMOVQ K1, K2;KMOVQ K1, K1;KMOVQ K1, K1;KMOVQ K1, K1;KMOVQ K1, K1
• Chain latency: 5
• Code:

     0:	62 f2 7d 29 8f d1    	vpshufbitqmb k2{k1},ymm0,ymm1
     6:	c4 e1 f8 90 ca       	kmovq  k1,k2
     b:	c4 e1 f8 90 c9       	kmovq  k1,k1
    10:	c4 e1 f8 90 c9       	kmovq  k1,k1
    15:	c4 e1 f8 90 c9       	kmovq  k1,k1
    1a:	c4 e1 f8 90 c9       	kmovq  k1,k1
  

• Init:

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

• Show nanoBench command
• Results:
  • MPERF: 7.41
  • APERF: 11.0
  • UOPS: 6.0

Experiment 1

• Instruction: VPSHUFBITQMB K1 {K2}, YMM0, YMM1
• Chain instruction: VMOVUPS YMM0 {K1}, YMM14
• Chain latency: 2
• Code:

     0:	62 f2 7d 2a 8f c9    	vpshufbitqmb k1{k2},ymm0,ymm1
     6:	62 d1 7c 29 10 c6    	vmovups ymm0{k1},ymm14
  

• Init:

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

• Show nanoBench command
• Results:
  • MPERF: 3.93
  • APERF: 5.99
  • UOPS: 2.0

Experiment 1

• Instruction: VPSHUFBITQMB K1 {K2}, YMM1, YMM0
• Chain instruction: VMOVUPS YMM0 {K1}, YMM14
• Chain latency: 2
• Code:

     0:	62 f2 75 2a 8f c8    	vpshufbitqmb k1{k2},ymm1,ymm0
     6:	62 d1 7c 29 10 c6    	vmovups ymm0{k1},ymm14
  

• Init:

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

• Show nanoBench command
• Results:
  • MPERF: 4.05
  • APERF: 6.0
  • UOPS: 2.0