SHORT Cycle Activities

EVENTSET
FIXC0 INSTR_RETIRED_ANY
FIXC1 CPU_CLK_UNHALTED_CORE
FIXC2 CPU_CLK_UNHALTED_REF
PMC0 CYCLE_ACTIVITY_STALLS_LDM_PENDING
PMC1 CYCLE_ACTIVITY_STALLS_L2_PENDING
PMC2 CYCLE_ACTIVITY_STALLS_L1D_PENDING
PMC3 CYCLE_ACTIVITY_CYCLES_NO_EXECUTE

METRICS
Runtime (RDTSC) [s] time
Runtime unhalted [s] FIXC1*inverseClock
Clock [MHz]  1.E-06*(FIXC1/FIXC2)/inverseClock
CPI  FIXC1/FIXC0
Cycles without execution [%] PMC3/FIXC1*100
Cycles with stalls due to L1D [%] PMC2/FIXC1*100
Cycles with stalls due to L2 [%] PMC1/FIXC1*100
Cycles with stalls due to LDM [%] PMC0/FIXC1*100

LONG
Formulas:
Cycles without execution [%] = 100.0 * CYCLE_ACTIVITY_CYCLES_NO_EXECUTE/CPU_CLK_UNHALTED_CORE
Cycles with stalls due to L1D [%] = 100.0 * CYCLE_ACTIVITY_STALLS_L1D_PENDING/CPU_CLK_UNHALTED_CORE
Cycles with stalls due to L2 [%] = 100.0 * CYCLE_ACTIVITY_STALLS_L2_PENDING/CPU_CLK_UNHALTED_CORE
Cycles with stalls due to LDM [%] = 100.0 * CYCLE_ACTIVITY_STALLS_LDM_PENDING/CPU_CLK_UNHALTED_CORE
--
This performance group measures the execution stalls due to load misses in
the L1D and L2 cache. The group measures only the cycles where the CPU cannot
execute any instruction, it does not measure the duration of stalls in the cache
layers.
