==3030== Cachegrind, a cache and branch-prediction profiler
==3030== Copyright (C) 2002-2017, and GNU GPL'd, by Nicholas Nethercote et al.
==3030== Using Valgrind-3.19.0 and LibVEX; rerun with -h for copyright info
==3030== Command: _bin/cxx-opt-sh/osh -n --ast-format none benchmarks/testdata/t9300-fast-import.sh
==3030== Parent PID: 3023
==3030== 
--3030-- warning: L3 cache found, using its data for the LL simulation.
==3030== 
==3030== I   refs:      11,866,632
==3030== I1  misses:        29,757
==3030== LLi misses:         7,801
==3030== I1  miss rate:       0.25%
==3030== LLi miss rate:       0.07%
==3030== 
==3030== D   refs:       5,171,658  (3,105,358 rd   + 2,066,300 wr)
==3030== D1  misses:        73,056  (   37,494 rd   +    35,562 wr)
==3030== LLd misses:        45,247  (   12,758 rd   +    32,489 wr)
==3030== D1  miss rate:        1.4% (      1.2%     +       1.7%  )
==3030== LLd miss rate:        0.9% (      0.4%     +       1.6%  )
==3030== 
==3030== LL refs:          102,813  (   67,251 rd   +    35,562 wr)
==3030== LL misses:         53,048  (   20,559 rd   +    32,489 wr)
==3030== LL miss rate:         0.3% (      0.1%     +       1.6%  )