XCHG (R64, R64) - Latency

Experiment 1

• Instruction: XCHG R8, R9
• Chain instructions: MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D
• Chain latency: 5
• Code:

     0:	4d 87 c8             	xchg   r8,r9
     3:	4d 63 c0             	movsxd r8,r8d
     6:	4d 63 c0             	movsxd r8,r8d
     9:	4d 63 c0             	movsxd r8,r8d
     c:	4d 63 c0             	movsxd r8,r8d
     f:	4d 63 c0             	movsxd r8,r8d
  

• Show nanoBench command
• Results:
  • MPERF: 2.28
  • APERF: 2.78
  • UOPS: 7.0

Experiment 2

• Instruction: XCHG R8, R9
• Code:

     0:	4d 87 c8             	xchg   r8,r9
  

• Show nanoBench command
• Results:
  • MPERF: 0.24
  • APERF: 0.33
  • UOPS: 2.0

Experiment 3 (with dependency-breaking instructions)

• Instruction: XCHG R8, R9
• Chain instruction: MOV R9, 0
• Code:

     0:	4d 87 c8             	xchg   r8,r9
     3:	49 c7 c1 00 00 00 00 	mov    r9,0x0
  

• Show nanoBench command
• Results:
  • MPERF: 0.6
  • APERF: 0.62
  • UOPS: 3.0

Experiment 1

• Instruction: XCHG R8, R9
• Chain instructions: MOVSX R8, R9D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D
• Chain latency: 5
• Code:

     0:	4d 87 c8             	xchg   r8,r9
     3:	4d 63 c1             	movsxd r8,r9d
     6:	4d 63 c0             	movsxd r8,r8d
     9:	4d 63 c0             	movsxd r8,r8d
     c:	4d 63 c0             	movsxd r8,r8d
     f:	4d 63 c0             	movsxd r8,r8d
  

• Show nanoBench command
• Results:
  • MPERF: 4.32
  • APERF: 5.0
  • UOPS: 7.0

Experiment 1

• Instruction: XCHG R9, R8
• Chain instructions: MOVSX R8, R9D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D
• Chain latency: 5
• Code:

     0:	4d 87 c1             	xchg   r9,r8
     3:	4d 63 c1             	movsxd r8,r9d
     6:	4d 63 c0             	movsxd r8,r8d
     9:	4d 63 c0             	movsxd r8,r8d
     c:	4d 63 c0             	movsxd r8,r8d
     f:	4d 63 c0             	movsxd r8,r8d
  

• Show nanoBench command
• Results:
  • MPERF: 4.02
  • APERF: 5.0
  • UOPS: 7.0

Experiment 1

• Instruction: XCHG R9, R8
• Chain instructions: MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D;MOVSX R8, R8D
• Chain latency: 5
• Code:

     0:	4d 87 c1             	xchg   r9,r8
     3:	4d 63 c0             	movsxd r8,r8d
     6:	4d 63 c0             	movsxd r8,r8d
     9:	4d 63 c0             	movsxd r8,r8d
     c:	4d 63 c0             	movsxd r8,r8d
     f:	4d 63 c0             	movsxd r8,r8d
  

• Show nanoBench command
• Results:
  • MPERF: 2.28
  • APERF: 2.78
  • UOPS: 7.0

Experiment 2

• Instruction: XCHG R9, R8
• Code:

     0:	4d 87 c1             	xchg   r9,r8
  

• Show nanoBench command
• Results:
  • MPERF: 0.3
  • APERF: 0.33
  • UOPS: 2.0

Experiment 3 (with dependency-breaking instructions)

• Instruction: XCHG R9, R8
• Chain instruction: MOV R9, 0
• Code:

     0:	4d 87 c1             	xchg   r9,r8
     3:	49 c7 c1 00 00 00 00 	mov    r9,0x0
  

• Show nanoBench command
• Results:
  • MPERF: 0.54
  • APERF: 0.62
  • UOPS: 3.0