Index

-calipers option
Collecting Data for Part of a Program

-gprof
example
Example of prof –gprof

_RLD_LIST variable
Effects of ssrun

_SPEEDSHOP_CALIPER_POINT_SIG variable
Using Calipers
User Environment Variables
Setting Calipers With Signals

_SPEEDSHOP_DEBUG_NO_SIG_TRAPS variable
Expert-Mode Environment Variables

_SPEEDSHOP_DEBUG_NO_STACK_UNWIND variable
Expert-Mode Environment Variables

_SPEEDSHOP_EXPERIMENT_TYPE variable
Effects of ssrun

_SPEEDSHOP_EXPERIMENT_TYPE variable
Expert-Mode Environment Variables

_SPEEDSHOP_FILE_BUFFER_LENGTH variable
Expert-Mode Environment Variables

_SPEEDSHOP_HWC_COUNTER_NUMBER variable
User Environment Variables
prof_hwc

_SPEEDSHOP_HWC_COUNTER_OVERFLOW variable
User Environment Variables
prof_hwc

_SPEEDSHOP_INIT_DEFERRED_SIGNAL variable
Expert-Mode Environment Variables

_SPEEDSHOP_MARCHING_ORDERS variable
Expert-Mode Environment Variables
Effects of ssrun

_SPEEDSHOP_OUTPUT_DIRECTORY variable
User Environment Variables

_SPEEDSHOP_OUTPUT_FILENAME variable
User Environment Variables

_SPEEDSHOP_OUTPUT_NOCOMPRESS variable
User Environment Variables

_SPEEDSHOP_REUSE_FILE_DESCRIPTORS variable
User Environment Variables

_SPEEDSHOP_SAMPLING_MODE variable
Expert-Mode Environment Variables

_SPEEDSHOP_SBRK_BUFFER_LENGTH variable
Expert-Mode Environment Variables

_SPEEDSHOP_SILENT variable
User Environment Variables

_SPEEDSHOP_TARGET_FILE variable
Effects of ssrun

_SPEEDSHOP_TRACE_EXEC variable
Process Tracking Environment Variables

_SPEEDSHOP_TRACE_FORK variable
Process Tracking Environment Variables

_SPEEDSHOP_TRACE_FORK_TO_EXEC variable
Process Tracking Environment Variables

_SPEEDSHOP_TRACE_SPROC variable
Process Tracking Environment Variables

_SPEEDSHOP_TRACE_SYSTEM variable
Process Tracking Environment Variables

_SPEEDSHOP_VERBOSE variable
User Environment Variables

API
API
setting calipers
Collecting Data for Part of a Program

basic block counting
ideal Experiment
ideal Experiment
ideal Experiment
overview
Experiment Types

C
examples
Tutorial for C Users

calipers
Building Your Executable
Using Calipers
Collecting Data for Part of a Program
and prof
Using the -calipers Option
automatic
Using Calipers
sample traps
Using Calipers
Setting Calipers With a Debugger
ssrt_caliper_point
Setting Calipers With ssrt_caliper_point
Using Calipers
using signals
Using Calipers
Setting Calipers With Signals
using the debugger
Setting Calipers With a Debugger
Using Calipers

call stack profiling
usertime Experiment
usertime Experiment
usertime Experiment

compiler feedback files
Generating Compiler Feedback Files

compiler optimization restrictions
Building Your Executable

cord
Dumping Compiler Feedback Files
Generating Compiler Feedback Files

.Counts file
Obtaining Basic Block Counts

CPU-bound processes
CPU-Bound Processes

cy_hwc experiment
[f]cy_hwc

data display anomalies
Special Information for MP Fortran Programs

dc_hwc experiment
[f]dc_hwc

debugger
setting calipers
Setting Calipers With a Debugger
Using Calipers
Collecting Data for Part of a Program
using ssrun
Using ssrun With a Debugger

dsc_hwc experiment
[f]dsc_hwc

DSOs
Supported Programming Models and Languages

environment variables
_SPEEDSHOP_MARCHING_ORDERS
Effects of ssrun
_SPEEDSHOP_TARGET_FILE
Effects of ssrun
LD_LIBRARY_PATH
Tutorial Setup

environment variables
_RLD_LIST
Effects of ssrun
_SPEEDSHOP_CALIPER_POINT_SIG
Using Calipers
Setting Calipers With Signals
User Environment Variables
_SPEEDSHOP_DEBUG_NO_SIG_TRAPS
Expert-Mode Environment Variables
_SPEEDSHOP_DEBUG_NO_STACK_UNWIND
Expert-Mode Environment Variables
_SPEEDSHOP_EXPERIMENT_TYPE
Effects of ssrun
Expert-Mode Environment Variables
_SPEEDSHOP_FILE_BUFFER_LENGTH
Expert-Mode Environment Variables
_SPEEDSHOP_HWC_COUNTER_NUMBER
User Environment Variables
_SPEEDSHOP_HWC_COUNTER_NUMBER
prof_hwc
_SPEEDSHOP_HWC_COUNTER_OVERFLOW
User Environment Variables
_SPEEDSHOP_HWC_COUNTER_OVERFLOW
prof_hwc
_SPEEDSHOP_INIT_DEFERRED_SIGNAL
Expert-Mode Environment Variables
_SPEEDSHOP_MARCHING_ORDERS
Expert-Mode Environment Variables
_SPEEDSHOP_OUTPUT_DIRECTORY
User Environment Variables
_SPEEDSHOP_OUTPUT_FILENAME
User Environment Variables
_SPEEDSHOP_OUTPUT_NOCOMPRESS
User Environment Variables
_SPEEDSHOP_REUSE_FILE_DESCRIPTORS
User Environment Variables
_SPEEDSHOP_SAMPLING_MODE
Expert-Mode Environment Variables
_SPEEDSHOP_SBRK_BUFFER_LENGTH
Expert-Mode Environment Variables
_SPEEDSHOP_SILENT
User Environment Variables
_SPEEDSHOP_TRACE_EXEC
Process Tracking Environment Variables
_SPEEDSHOP_TRACE_FORK
Process Tracking Environment Variables
_SPEEDSHOP_TRACE_FORK_TO_EXEC
Process Tracking Environment Variables
_SPEEDSHOP_TRACE_SPROC
Process Tracking Environment Variables
_SPEEDSHOP_TRACE_SYSTEM
Process Tracking Environment Variables
_SPEEDSHOP_VERBOSE
User Environment Variables

examples
C
Tutorial for C Users
Fortran
Tutorial for Fortran Users

exec
Supported Programming Models and Languages

executables
calculating a working set
Calculating the Working Set of a Program
stripped
Building Your Executable

experiments
fpe trace
fpe Trace

experiment data
Using ssrun and prof to Gather and Analyze Performance Data
controlling output file
Setting Up Output Directories and Files
file format
Experiment File Format
filenames
Setting Up Output Directories and Files

experiments
choosing
Selecting an Experiment
Using ssrun and prof to Gather and Analyze Performance Data
cy_hwc
[f]cy_hwc
dc_hwc
[f]dc_hwc
dsc_hwc
[f]dsc_hwc
fpe
FPE Trace Reports
fpe trace
Using ssrun and prof to Gather and Analyze Performance Data
fpe Trace
gfp_hwc
[f]gfp_hwc
gi_hwc
[f]gi_hwc
hardware counter
Hardware Counter Experiments
Hardware Counter Experiment Reports
Hardware Counter Experiment
Hardware Counter Experiment
hardware counters
Using ssrun and prof to Gather and Analyze Performance Data
ic_hwc
[f]ic_hwc
ideal
ideal Experiment
ideal Experiment
ideal Experiment
ideal Experiment Reports
isc_hwc
[f]isc_hwc
pcsamp
pcsamp Experiment
pcsamp
Using ssrun and prof to Gather and Analyze Performance Data
pcsamp Experiment
pcsamp Experiment
pcsamp Experiment Reports
prof_hwc
prof_hwc
tlb_hwc
[f]tlb_hwc
usertime
usertime Experiment Reports
usertime Experiment
Using ssrun and prof to Gather and Analyze Performance Data
usertime Experiment
usertime Experiment

fbdump
Dumping Compiler Feedback Files
overview
Additional Commands

files
compiler feedback
Dumping Compiler Feedback Files
performance data
Using ssrun and prof to Gather and Analyze Performance Data
Dumping Performance Data Files
format
Experiment File Format

floating-point exception trace
Using ssrun and prof to Gather and Analyze Performance Data
overview
Experiment Types

floating-point exceptions
fpe Trace
fpe Trace

fork
Supported Programming Models and Languages

Fortran
examples
Tutorial for Fortran Users
limitations
Special Information for MP Fortran Programs

fpcsampx experiment
pcsamp Experiment

fpe trace experiment
FPE Trace Reports
fpe Trace
Using ssrun and prof to Gather and Analyze Performance Data
overview
Experiment Types
tutorial
fpe Trace

generic program
Tutorial Overview

gfp_hwc experiment
[f]gfp_hwc

gi_hwc experiment
[f]gi_hwc

hardware counter experiment
Hardware Counter Experiment Reports

hardware counter experiments
Using ssrun and prof to Gather and Analyze Performance Data
Hardware Counter Experiments
overview
Experiment Types
tutorial
Hardware Counter Experiment
Hardware Counter Experiment

hardware counter numbers
Hardware Counter Numbers

hardware counter overflows
Hardware Counter Experiment
Hardware Counter Experiments
Hardware Counter Experiment

hwc experiments
Using ssrun and prof to Gather and Analyze Performance Data
Hardware Counter Experiments
overview
Experiment Types

I/O-bound processes
I/O-Bound Processes

ic_hwc experiment
[f]ic_hwc

ideal experiment
ideal Experiment
ideal Experiment Reports
effects
Effects of ssrun
overview
Experiment Types
tutorial
ideal Experiment
ideal Experiment

isc_hwc experiment
[f]isc_hwc

LD_LIBRARY_PATH variable
Tutorial Setup
Obtaining Basic Block Counts

libfpe_ss.so
SpeedShop Libraries

libmalloc_ss.so
SpeedShop Libraries

libraries
libfpe_ss.so
SpeedShop Libraries
libmalloc_ss.so
SpeedShop Libraries
libss.so
Effects of ssrun
SpeedShop Libraries
libssrt.so
SpeedShop Libraries
Effects of ssrun
linking in SpeedShop
Setting Calipers With ssrt_caliper_point

libss.so
Setting Calipers With ssrt_caliper_point
Effects of ssrun
SpeedShop Libraries

libssrt.so
Setting Calipers With ssrt_caliper_point
SpeedShop Libraries
Effects of ssrun

linpack benchmark
Tutorial Overview

locking memory
Using the squeeze Command

machine resource usage
Collecting Data on Machine Resource Usage

memory
locking
Using the squeeze Command

memory-bound processes
Memory-Bound Processes

message-passing paradigms
Supported Programming Models and Languages

MP Fortran limitations
Special Information for MP Fortran Programs

MPI
Supported Programming Models and Languages
with ssrun
Running Experiments on MPI Programs

multi-processor executables
Special Information for MP Fortran Programs
Supported Programming Models and Languages
profiling
Generating Reports for Multiprocessed Executables

paging behavior
Using the thrash Command

PC sampling
pcsamp Experiment
tutorial
pcsamp Experiment
pcsamp Experiment

pcsamp experiment
pcsamp Experiment Reports
pcsamp Experiment
Using ssrun and prof to Gather and Analyze Performance Data
example
Example Using the pcsampx Experiment
overview
Experiment Types
tutorial
pcsamp Experiment
pcsamp Experiment

perfex
Two Tools for Hardware Counter Experiments

performance analysis
phases
Using SpeedShop Tools for Performance Analysis
theory
Sources of Performance Problems

performance data files
dumping
Dumping Performance Data Files

performance problems
Sources of Performance Problems
Using ssrun and prof to Gather and Analyze Performance Data
Bugs
Bugs
CPU
CPU-Bound Processes
I/O
I/O-Bound Processes
memory
Memory-Bound Processes

pixie
ideal Experiment
Using pixie
–autopixie option
pixie Options
–verbose option
pixie Options
and prof –heavy example
Example Using prof –heavy
and prof –i example
Example Using prof –invocations
command option
Using pixie
command syntax
Using pixie
.Counts file
Obtaining Basic Block Counts
examples
Obtaining Basic Block Counts
output size
Obtaining Basic Block Counts
overview
Additional Commands
restricting output
Obtaining Basic Block Counts
setting search path
Obtaining Basic Block Counts

prof
-calipers example
Using the -calipers Option
-dis example
Using the -dis Option
-gprof example
Using the -gprof Option
-S example
Using the -S Option

processes
forking
Supported Programming Models and Languages

prof
–heavy example
Example Using prof –heavy
–invocations example
Example Using prof –invocations
-calipers option
Collecting Data for Part of a Program
-gprof example
ideal Experiment Reports
Also see profiling
prof Syntax
compiler feedback
Dumping Compiler Feedback Files
options
prof Options
output
prof Output
overview
Using ssrun and prof to Gather and Analyze Performance Data
Main Commands
steps
Using ssrun and prof to Gather and Analyze Performance Data
syntax
prof Syntax
using with pixie
Using prof to Generate Performance Reports
using with ssrun
Using prof to Generate Performance Reports

prof_hwc experiment
prof_hwc

profiles
interpreting
Interpreting Reports

profiling
–clock option
prof Options
–dis option
prof Options
–dso option
prof Options
–dsolist option
prof Options
–exclude option
prof Options
–feedback option
prof Options
–heavy option
prof Options
example
Example Using prof –heavy
–invocations option
prof Options
example
Example Using prof –invocations
–lines option
prof Options
–only option
prof Options
–procedures option
prof Options
–quit option
Example Using prof –quit
Obtaining Basic Block Counts
prof Options
–S option
prof Options
–zero option
prof Options
-calipers option
Using the -calipers Option
-dis option
Using the -dis Option
-gprof option
prof Options
Using the -gprof Option
-S option
Using the -S Option
command syntax
prof Syntax
fpe trace experiment
FPE Trace Reports
hardware counter experiments
Hardware Counter Experiment Reports
ideal experiment
ideal Experiment Reports
inclusive basic block counts
ideal Experiment Reports
machine scheduler option
Generating Reports for Different Machine Types
multiprocessor executables
Generating Reports for Multiprocessed Executables
pcsamp experiment
pcsamp Experiment Reports
procedure invocation example
Examples of Basic Block Counting
processor scheduler option option
prof Options
usertime experiment
usertime Experiment Reports

program counter sampling
pcsamp Experiment

programs
calculating a working set
Calculating the Working Set of a Program
stripped
Building Your Executable

pthreads
Supported Programming Models and Languages
and ssrun
Running Experiments on Programs Using Pthreads

rearranging procedures
Generating Compiler Feedback Files

reports
for different machine models
Generating Reports for Different Machine Types
fpe trace experiment
FPE Trace Reports
hardware counter experiments
Hardware Counter Experiment Reports
ideal experiment
ideal Experiment Reports
interpreting
Interpreting Reports
pcsamp experiment
pcsamp Experiment Reports
usertime experiment
usertime Experiment Reports
using calipers
Using the -calipers Option

rld
search path
Obtaining Basic Block Counts

search path
rld
Obtaining Basic Block Counts

setting calipers
Using Calipers
Collecting Data for Part of a Program

shared libraries
Supported Programming Models and Languages

signals
setting calipers
Setting Calipers With Signals
Collecting Data for Part of a Program
Using Calipers

SpeedShop
overview
SpeedShop Tools

SpeedShop API
API

SpeedShop demo
generic
Tutorial Overview
linpack
Tutorial Overview

SpeedShop libraries
linking
Setting Calipers With ssrt_caliper_point

SpeedShop libraries
Effects of ssrun
libfpe_ss.so
SpeedShop Libraries
libmalloc_ss.so
SpeedShop Libraries
libss.so
SpeedShop Libraries
libssrt.so
SpeedShop Libraries

sproc
Supported Programming Models and Languages

squeeze
Using the squeeze Command
calculating a working set
Calculating the Working Set of a Program
overview
Additional Commands

ssdump
Dumping Performance Data Files

ssrt_caliper_point
Using Calipers
Setting Calipers With ssrt_caliper_point
API
executable requirements
Building Your Executable

ssrun
-v option example
Example Using the -v Option
effects
Effects of ssrun
examples
ssrun Examples
flags
ssrun Syntax
MPI programs
Running Experiments on MPI Programs
overview
Using ssrun and prof to Gather and Analyze Performance Data
Main Commands
pthreads programs
Running Experiments on Programs Using Pthreads
restrictions
Building Your Executable
setup
Building Your Executable
steps
Using ssrun and prof to Gather and Analyze Performance Data
syntax
ssrun Syntax
using a debugger
Using ssrun With a Debugger

ssusage
calculating a working set
Calculating the Working Set of a Program
overview
Main Commands

statistical call stack profiling
overview
Experiment Types

statistical hardware counter sampling
overview
Experiment Types

statistical PC sampling
overview
Experiment Types

stripped executables
Building Your Executable

system
Supported Programming Models and Languages

thrash
Using the thrash Command
calculating a working set
Calculating the Working Set of a Program
overview
Additional Commands

tlb_hwc experiment
[f]tlb_hwc

tracing floating-point exceptions
Using ssrun and prof to Gather and Analyze Performance Data

tutorial
C
Tutorial for C Users
Fortran
Tutorial for Fortran Users

usertime experiment
usertime Experiment
Using ssrun and prof to Gather and Analyze Performance Data
usertime Experiment Reports
overview
Experiment Types
restrictions
Building Your Executable
tutorial
usertime Experiment
usertime Experiment

working set
Calculating the Working Set of a Program