==3294== Cachegrind, a cache and branch-prediction profiler
==3294== Copyright (C) 2002-2017, and GNU GPL'd, by Nicholas Nethercote et al.
==3294== Using Valgrind-3.14.0 and LibVEX; rerun with -h for copyright info
==3294== Command: mksh -n benchmarks/testdata/test-cmd-util.sh
==3294== Parent PID: 3289
==3294== 
--3294-- warning: L3 cache found, using its data for the LL simulation.
==3294== 
==3294== I   refs:      17,050,251
==3294== I1  misses:        32,934
==3294== LLi misses:         2,134
==3294== I1  miss rate:       0.19%
==3294== LLi miss rate:       0.01%
==3294== 
==3294== D   refs:       6,109,154  (3,895,184 rd   + 2,213,970 wr)
==3294== D1  misses:        48,171  (   38,547 rd   +     9,624 wr)
==3294== LLd misses:         9,049  (    3,039 rd   +     6,010 wr)
==3294== D1  miss rate:        0.8% (      1.0%     +       0.4%  )
==3294== LLd miss rate:        0.1% (      0.1%     +       0.3%  )
==3294== 
==3294== LL refs:           81,105  (   71,481 rd   +     9,624 wr)
==3294== LL misses:         11,183  (    5,173 rd   +     6,010 wr)
==3294== LL miss rate:         0.0% (      0.0%     +       0.3%  )