| e806829c9a1f8e16170d70b8b8190322794b048c |
|
24-Aug-2016 |
Mathias Agopian <mathias@google.com> |
report MPKI instead of hit-rate
MPKI (or misses per 1000 instructions) is better
for comparing CPUs, as hit-rate somewhat
meaningless (different CPUs make different
amount of requests to the same cache-line).
Also changed the —affinity option to take the
CPU number instead of a big|little string
Change-Id: I57fce77e18d382fe8d4f0efd9cc586d50beb9432
/system/extras/tests/icachetest/icache_main.cpp
|