MUL (R64) - Throughput and Uops
With unroll_count=500 and no inner loop
Code:
0: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 1.0
Core cycles: 7.0
Reference cycles: 3.48
UOPS_RETIRED.ALL: 3.0
UOPS_MS: 3.0
With loop_count=1000 and unroll_count=10
Code:
0: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 1.2
Core cycles: 7.2
Reference cycles: 3.6
UOPS_RETIRED.ALL: 3.2
UOPS_MS: 3.0
With loop_count=100 and unroll_count=100
Code:
0: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 1.02
Core cycles: 7.02
Reference cycles: 3.51
UOPS_RETIRED.ALL: 3.02
UOPS_MS: 3.0
With additional dependency-breaking instructions
With unroll_count=500 and no inner loop
Code:
0: 48 31 c0 xor rax,rax 3: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 2.0
Core cycles: 8.0
Reference cycles: 4.0
UOPS_RETIRED.ALL: 4.0
UOPS_MS: 3.0
With loop_count=1000 and unroll_count=10
Code:
0: 48 31 c0 xor rax,rax 3: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 2.2
Core cycles: 8.1
Reference cycles: 4.05
UOPS_RETIRED.ALL: 4.2
UOPS_MS: 3.0
With loop_count=100 and unroll_count=100
Code:
0: 48 31 c0 xor rax,rax 3: 49 f7 e0 mul r8
Show nanoBench command
Results:
Instructions retired: 2.02
Core cycles: 8.01
Reference cycles: 4.01
UOPS_RETIRED.ALL: 4.02
UOPS_MS: 3.0