perf-stat

Collect hardware performance counters using perf stat.

Runs perf stat on cores 0 to N_CORES-1 for PERF_TIME seconds, collecting
events such as LLC misses, L1 cache misses, branch misses, TLB misses, and
instructions per cycle (IPC). Computes and outputs miss ratios.

Use the tag "perf_all" to enable a more set of hardware events than the default events.

Usage:
  %import@dut perf-stat N_CORES=4 PERF_TIME=5

See also :doc:`perf-class <modules/perf-class>`

// Perf stat