valgrind

$ valgrind --tool=cachegrind ./test
==11710== Cachegrind, a cache and branch-prediction profiler
==11710== Copyright (C) 2002-2013, and GNU GPL'd, by Nicholas Nethercote et al.
==11710== Using Valgrind-3.10.0.SVN and LibVEX; rerun with -h for copyright info
==11710== Command: ./test
==11710==
--11710-- warning: L3 cache found, using its data for the LL simulation.
--11710-- warning: pretending that LL cache has associativity 30 instead of actual 20
==11710==
==11710== I   refs:      163,367
==11710== I1  misses:        869
==11710== LLi misses:        861
==11710== I1  miss rate:    0.53%
==11710== LLi miss rate:    0.52%
==11710==
==11710== D   refs:       54,700  (40,915 rd   + 13,785 wr)
==11710== D1  misses:      2,940  ( 2,391 rd   +    549 wr)
==11710== LLd misses:      2,406  ( 1,904 rd   +    502 wr)
==11710== D1  miss rate:     5.3% (   5.8%     +    3.9%  )
==11710== LLd miss rate:     4.3% (   4.6%     +    3.6%  )
==11710==
==11710== LL refs:         3,809  ( 3,260 rd   +    549 wr)
==11710== LL misses:       3,267  ( 2,765 rd   +    502 wr)
==11710== LL miss rate:      1.4% (   1.3%     +    3.6%  )