HP Caliper User Guide Release 5.5 (5900-2351, August 2012)

be obtained from lower level caches or memory, leading
to extra cycles.
CYC_BE_DET_REPLAY.ANY This is the number of cycles lost in replay in the DET stage.
CYC_BE_EXE_REPLAY.ANY This is the number of cycles lost in replay in the EXE stage.
CYC_BE_WB2_REPLAY.ANY This is the number of cycles lost in replay in the WB2 stage.
% Unstalled execution (higher is
better)
Percentage of unstalled cycles with respect to total number
of elapsed CPU operating cycles.
% Cycles lost due to branch
misprediction or interruption flush
stalls (lower is better)
The percentage of cycles lost due to flushes in the WB2
stage with respect to the the total number of elapsed CPU
operating cycles.
% Cycles lost due to issue bubbles
(lower is better)
The percentage of cycles lost due to stalls in IBD stage with
respect to total number of elapsed CPU operating cycles.
% Cycles lost due to DET replay
(lower is better)
The percentage of cycles lost due to replays in DET stage
with respect to total number of elapsed CPU operating
cycles.
% Cycles lost due to EXE replay
(lower is better)
The percentage of cycles lost due to replays in EXE stage
with respect to total number of elapsed CPU operating
cycles.
% Cycles lost due to WB2 replay
(lower is better)
The percentage of cycles lost due to replays in WB2 stage
with respect to total number of elapsed CPU operating
cycles.
% of Cycles lost due to GR load
penalties
Percentage of cycles lost due to GR load dependency stalls
(stalls in IBD, RAW and WAW hazards in EXE and DET
replays) with respect to total number of elapsed CPU
operating cycles.
% Effective instructions retired The total number of instructions retired without considering
the nop's and instructions with predicate disabled.
% Raw CPI (lower is better) The cycles per instruction, including nop and predicated off
instructions.
% Raw IPC (higher is better) The instructions per cycle, including nop and predicated off
instructions.
% Effective CPI (lower is better) The cycles per effective instruction, excluding nop and
predicated off instructions.
% Effective IPC (higher is better) The effective instructions per cycle, excluding nop and
predicated off instructions.
% Core cycles due to this thread This indicates the percentage of available processor cycles
that the measured process consumed. The other processor
cycles were consumed by other process(es) running in the
core's other hyperthread or were lost to HyperThreading
overhead.
fprof Measurement Metrics
See Table 26 (page 212).
In this table, “program object” refers to any of the following:
• Thread
• Load module
• Function
• Source statement
• Instruction bundle
fprof Measurement Report Description 211