HP Caliper 5.3 User Guide (5900-1558, February 2011)

itlb Measurement Report Description
With the itlb measurement, produced by the itlb measurement configuration file, HP Caliper
measures and reports two levels of information:
Exact counts of instruction translation lookaside buffer (TLB) metrics summed across the entire
run of an application
Sampled instruction TLB metrics that are associated with particular locations in the application
The report shows masured data by thread, load module, function, statement, and cache line.
Command-line options allow you to control the amount of data reported, how the data is sorted,
and the number of statements and instructions reported for each sampled program location.
Example Command Line for Text Report
$ caliper itlb -o reports/itlbm.txt ./matmul
Example Command Line for CSV Report
$ caliper itlb --csv csvout ./matmul
itlb Metrics Summed for Entire Run
This section describes the metrics summed over the entire run of your application under HP Caliper.
Metrics for Integrity Servers Itanium 2 Systems
L1I_READS The number of data memory read references issued into the
memory pipeline that are serviced by L1D (only integer
loads), RSE loads, L1-hinted loads (L1D returns data if it hits
in L1D but does not do a fill), and check loads (ld.c).
Non-cacheable reads, VHPT loads, semaphores,
floating-point loads, and lfetch instructions are not counted
here because L1D does not handle these. The count includes
wrong path operations but excludes predicated off
operations.
ITLB_MISSES_FETCH.L1ITLB Number of L1 instruction TLB misses for demand fetch.
Misses are counted even if the L1 instruction TLB is not
updated for an access (non-cacheable/nat page/not present
page/faulting/some flushed), it will be counted here.
ITLB_MISSES_FETCH.L2ITLB Number of L1 instruction TLB misses that also missed in the
L2 instruction TLB.
Total L1 Instruction TLB References Total number of L1 instruction TLB references.
L1 Instruction TLB Miss Ratio Ratio of L1 instruction TLB misses to Total L1 instruction TLB
references.
Metrics for Integrity Servers Dual-Core Itanium 2 and Itanium 9300 Quad-Core Processor Systems
BACK_END_BUBBLE.ALL Number of cycles when the back end of the pipeline was
stalled. This is the number of cycles lost (stall cycles) due to
any of five possible events (FPU/L1D, RSE, EXE,
branch/exception, or the front end).
BACK_END_BUBBLE.FE Full Pipe Bubbles in Main Pipe due to front end. This is the
number of cycles lost (stall cycles) due to ICACHE ITLB and
branch execution stalls.
BE_LOST_BW_DUE_TO_FE.ALL Number of invalid bundles at the exit from Instruction Buffer
only if the back end is not stalled for other reasons.
208 Descriptions of Measurement Reports