[tom@qemu]$~/bin/perf stat -e L1-dcache-loads,L1-dcache-load-misses,cpu-cycles,instructions,branch-instructions,branch-misses,branch-loads,branch-load-misses,dTLB-loads,dTLB-load-misses ./qemu-img co_bench -c 5000000 -s 512
co_bench: iterations 5000000, bypass: no, read_size: 512
	total time: 14469ms,   345K ops

 Performance counter stats for './qemu-img co_bench -c 5000000 -s 512':

    13,684,562,790      L1-dcache-loads                                              [39.99%]
        45,768,542      L1-dcache-load-misses     #    0.33% of all L1-dcache hits   [40.29%]
    40,174,086,427      cpu-cycles                [50.23%]
    48,394,067,389      instructions              #    1.20  insns per cycle         [60.20%]
     9,981,932,910      branch-instructions                                          [60.38%]
        73,464,674      branch-misses             #    0.74% of all branches         [60.30%]
     9,979,424,700      branch-loads                                                 [60.01%]
     7,186,019,818      branch-load-misses                                           [59.71%]
    13,682,610,805      dTLB-loads                                                   [39.62%]
        35,709,691      dTLB-load-misses          #    0.26% of all dTLB cache hits  [39.71%]

      14.471802631 seconds time elapsed

[tom@qemu]$
[tom@qemu]$~/bin/perf stat -e L1-dcache-loads,L1-dcache-load-misses,cpu-cycles,instructions,branch-instructions,branch-misses,branch-loads,branch-load-misses,dTLB-loads,dTLB-load-misses ./qemu-img co_bench -c 5000000 -s 512 -b
co_bench: iterations 5000000, bypass: yes, read_size: 512
	total time: 12091ms,   413K ops

 Performance counter stats for './qemu-img co_bench -c 5000000 -s 512 -b':

    11,796,271,295      L1-dcache-loads                                              [40.11%]
        38,402,609      L1-dcache-load-misses     #    0.33% of all L1-dcache hits   [40.06%]
    33,543,263,137      cpu-cycles                [50.01%]
    42,547,232,118      instructions              #    1.27  insns per cycle         [59.95%]
     8,633,815,806      branch-instructions                                          [59.95%]
        29,112,915      branch-misses             #    0.34% of all branches         [59.88%]
     8,635,488,297      branch-loads                                                 [59.89%]
     6,038,581,339      branch-load-misses                                           [59.96%]
    11,798,495,576      dTLB-loads                                                   [40.06%]
        47,069,986      dTLB-load-misses          #    0.40% of all dTLB cache hits  [40.12%]

      12.093324392 seconds time elapsed