Index

Symbols

--build-directories command line option --capture-environment-variables command line option --cuda-kernel-analysis command line option --cuda-transfer-analysis command line option --disable-metrics command line option --embed-compiler-remarks command line option --enable-metrics command line option --exclude-compiler-remarks command line option --export command line option --export-functions command line option --list-metrics command line option --no-mpi command line option

--no-queue command line option --perf-metrics command line option --profile command line option --queue command line option --report command line option --select-ranks command line option --spe command line option --start-after command line option --stop-after command line option --target-host command line option --view command line option

A

Accelerator Accelerator breakdown GPU utilization Mean GPU memory usage Peak GPU memory usage Align stacks AMD OpenCL AMD ROCm Known issues Application, [1] Arbitrary expressions and global variables Arm (AArch64) Array Auto Update Comparing elements across processes Data table Distributed Export Expression Filtering Multi-dimensional viewing Statistics Visualization

Array data Viewing Arrays Multi-dimensional Assembly debugging Breakpoints Toggling and viewing Attaching, [1], [2], [3], [4] Choose hosts Command line Hosts file Attaching to running programs AUTO_LAUNCH_TAG

B

Backtrace Breakpoints Conditional Default Deleting Focus Loading Pending Saving Setting Using source code viewer Using the Add Breakpoint window

Building applications, [1]

C

C++ STL support Caliper Capture Environment Variables command line option --build-directories --capture-environment-variables --cuda-kernel-analysis --cuda-transfer-analysis --disable-metrics --embed-compiler-remarks --enable-metrics --exclude-compiler-remarks --export --export-functions --list-metrics --no-mpi --no-queue --perf-metrics --profile --queue --report --select-ranks --spe --start-after --stop-after --target-host --view Compatibility Launch Compilers AMD Cray NVIDIA HPC OpenCL Complex numbers Configuring Remote hosts Core Files Core files CPU breakdown Memory accesses OpenMP code Scalar numeric ops Single core code Vector numeric ops Waiting for accelerators

CPU metrics breakdown Cycles per instruction L2 cache misses Stalled cycles Cray, [1] Compiling scalar programs Cray compiler environment Cray Native SLURM Cray XK6 Cross-process comparison Cross-thread comparison CSV performance reports CUDA Attaching Breakpoints Controlling GPU threads CUDA Fortran Debugging multiple CUDA processes GPU Debugging GPU device information GPU Language support Launching Licensing NVIDIA Preparing to debug Running Source code viewer Stepping Thread control Understanding kernel progress Current line Custom DCIM Custom gmetric Custom MPI scripts

D

Data Changing DCIM output Deadlock

Debugging Scalar Disassembler Duration

E

Editing source code, [1] Editing variables Enable and disable metrics Energy breakdown CPU Mean node power Peak node power System Energy metrics Requirements environment variable FORGE_CONFIG_DIR, [1], [2], [3] FORGE_DCIM_SCRIPT FORGE_DEBUG_HEURISTICS FORGE_DETECT_APRUN_VERSION FORGE_ENABLE_ALL_REPORTS_GENERATION FORGE_FORCE_LICENSE_FILE, [1], [2] FORGE_GMETRIC FORGE_INTEL_XE_THREAD_DETAILS, [1] FORGE_KEEP_OUTPUT_LINE_LENGTH, [1] FORGE_KEEP_OUTPUT_LINES, [1] FORGE_LICENSE_DIR, [1], [2] FORGE_LICENSE_FILE, [1], [2], [3] FORGE_MAC_INTERFACE FORGE_MAP_TO_DCIM FORGE_MPI_FINALIZE_TIMEOUT_MS FORGE_MPI_WRAPPER, [1], [2], [3] FORGE_MPIRUN, [1] FORGE_NEVER_FORWARD_STDIN FORGE_NO_APPLICATION_PROBE FORGE_NO_TIMEOUT FORGE_NOTES, [1] FORGE_PRESERVE_WRAPPER FORGE_PROCESS_TIMEOUT FORGE_SAMPLER_INTERVAL, [1], [2], [3], [4] FORGE_SAMPLER_INTERVAL_PER_THREAD, [1], [2], [3], [4], [5], [6] FORGE_SAMPLER_NO_TIME_MPI_CALLS FORGE_SAMPLER_NUM_SAMPLES, [1], [2] FORGE_SAMPLER_TRY_USE_SMAPS, [1], [2] FORGE_SYSTEM_CONFIG FORGE_USER_CONFIG FORGE_WRAPPER_COMPILE, [1] MPICC, [1], [2], [3]

Environment variables, [1], [2] Example Compiling Compiling Cray Generating a performance report Generating a performance report and .map file simultaneously Generating a performance report from an existing .map file Running Viewing/exporting a performance report whilst viewing an existing .map file Express Launch, [1], [2] Compatible MPIs Run dialog box Expression Changing language

F

Files Viewing multiple Find in Files Flexible Array Focus Breakpoints Code viewer Parallel stack view Playing Process group viewer Step threads together Stepping Focus control forge-probe FORGE_CONFIG_DIR, [1], [2] FORGE_FORCE_LICENSE_FILE, [1]

FORGE_INTEL_XE_THREAD_DETAILS FORGE_KEEP_OUTPUT_LINE_LENGTH FORGE_KEEP_OUTPUT_LINES FORGE_LICENSE_DIR, [1] FORGE_LICENSE_FILE, [1], [2] FORGE_MPI_WRAPPER, [1], [2] FORGE_MPIRUN FORGE_NOTES FORGE_SAMPLER_INTERVAL, [1], [2], [3] FORGE_SAMPLER_INTERVAL_PER_THREAD, [1], [2], [3], [4], [5] FORGE_SAMPLER_NUM_SAMPLES, [1] FORGE_SAMPLER_TRY_USE_SMAPS, [1] FORGE_WRAPPER_COMPILE Fortran Array Fortran intrinsics Fortran Modules Function Listing

G

Generating a report GNU/Linux systems Go To Line GPU Attaching, [1], [2] Breakpoints Controlling GPU threads CUDA Device information Examining GPU threads and data Intel Xe ROCm Selecting GPU threads Stepping Viewing GPU thread locations

GPU Debugging, [1], [2]

H

hipcc

HTML reports

I

I/O breakdown Effective process read rate Effective process write rate Lustre metrics Time in reads Time in writes I/O time icpx Intel MPI MPMD remote-exec Intel Xe Attaching, [1] Breakpoints Controlling GPU threads GPU device information GPU Thread Details Intel oneAPI Known issues, [1] Launching Licensing Preparing to debug Running Stepping

Intel Xeon Intel Xeon RAPL Interpreting

J

Job ID regular expression Job scheduling Job submission, [1] Cancelling Custom Regular expression, [1]

Jump To Line Double clicking

K

Known issues Cray Platform

L

Libraries Shared Libraries Licensing Single process license Single-process license Linaro DDT Getting started Getting Support Logbook Overview Program input and output Running a program Starting a program Starting a program from a job script Linaro MAP Environment variables Getting Support JSON JSON Activities JSON Categories JSON Example Overview Program details Program output Project files view Restricting output Saving output Standard error Standard output Starting from job script Time display mode

Linaro Performance Reports Getting Support Local variables Logbook Annotation Comparison window Linaro DDT Logbook Usage

M

MAC OS X Macros Manual launch Debugging multi-process non-MPI programs forge-client Manual process selection MAP file Memory breakdown Mean process memory usage Peak node memory usage Peak process memory usage Memory debugging Enabling Memory leak report Message queues Deadlock Interpreting Viewing Metrics Accelerator, [1] Accelerator breakdown Computation, [1] Compute CPU breakdown CPU metrics breakdown Cycles per instruction Effective process collective rate Effective process point-to-point rate Effective process read rate Effective process write rate Energy breakdown Energy CPU GPU Utilization I/O breakdown I/O time Input/Output L2 cache misses Lustre metrics Mean GPU memory usage Mean node power Mean process memory usage Memory accesses Memory breakdown MPI MPI breakdown MPI communication and waiting time OpenMP breakdown OpenMP code Peak GPU memory usage Peak node memory usage Peak node power Peak process memory usage Physical core utilization, [1] Scalar numeric ops Single core code Single-threaded computation time Stalled cycles Synchronization, [1] System System load, [1] Thread affinity breakdown Thread affinity max load Thread affinity migration opportunity Thread Affinity Utilization Threads breakdown Time in collective calls Time in point-to-point calls Time in reads Time in writes Vector numeric ops Waiting for accelerators

MPI Function Counters History/Logging MPI rank MPI Ranks Running MPI breakdown Effective process collective rate Effective process point-to-point rate Time in collective calls Time in point-to-point calls MPI communication and waiting time MPI job Attaching to a subset Automatic detection MPI_Init remote-exec MPICC, [1], [2] MPICH MPICH 3 MPMD remote-exec mpiicpx mpirun remote-exec MPMD Compatibility mode Intel MPI MPICH 3 remote-exec Running MPMD programs Compatibility mode Debugging Without Express Launch Multi-Dimensional Array Viewer (MDA)

N

Navigating through source code history Numactl DDT MAP Number bases Viewing

nvcc NVIDIA CUDA Known issues NVIDIA HPC

O

Obtaining Help Offline debugging HTML report Periodic snapshots Plain text report Reading a file for standard input Run-time job progress reporting Signal-triggered snapshots Using Writing a file from standard output Open MPI MPMD MPMD Compatibility mode OpenACC Cray NVIDIA HPC

OpenCL, [1] OpenGL OpenMP Debugging OMP_NUM_THREADS Running, [1] OpenMP breakdown Computation Physical core utilization Synchronization System load Output locations

P

Pending breakpoints Perf Metrics Performance reports Accelerator Threads breakdown Synchronization Plugins Enabling Installing Reference Supported Using Writing

Pointers Pretty printers Process details Process Group Viewer Process groups Deleting Detailed view Summary view PROCS_PER_NODE_TAG Profile a Python script Profiling, [1] Program part Python Profiling Python profiling known issues

Q

Queue submission Cancelling Queue submission via Express Launch Queue template syntax PROCS_PER_NODE_TAG

Queue template tags AUTO_LAUNCH_TAG Defining new tags Launching Specifying default options Using forge-mpirun

R

Raw command Raw Command Window Rebuilding applications, [1] Receive queue Registers Viewing remote-exec Required Report summary Compute Input/Output MPI Requirements Energy metrics Restarting ROCm, [1] Attaching, [1] Breakpoints Controlling GPU threads GPU device information GPU Language support Launching Licensing Preparing to debug Running Stepping

Run-time Job progress reporting Running MPMD Scalar Running a program Running programs Attaching Manual process selection

S

Saving output Scalar Debugging Running Scalar programs Search, [1] Security Shared HPC environments Vulnerability reporting Send queue Send signal Sending signals Session Saving Session menu Signal handling Custom Sending signals SIGUSR1 SIGUSR2 Single stepping Single-threaded computation time Source Code

Source code, [1] Application and external code split Committing, [1] Editing, [1] Find in Files Missing files Project files Rebuilding, [1] Searching, [1] Viewing, [1] Sparkline Sparklines SPE Stack frame Stacks view Standard error Standard input, [1] Standard output Static analysis Step threads together Stop messages Stopping Suspending breakpoints Synchronizing processes

T

Textual performance reports Thread affinity breakdown Max Load Migration opportunity Utilization Threads breakdown Computation Physical core utilization System load

Tracepoints Output Setting

U

Unexpected queue

UPC Using custom MPI scripts

V

Variables Searching, [1] Version control Breakpoints and tracepoints

Version control information Viewing multiple files Viewing stacks Overview Viewing stacks in parallel

W

Warning Symbols Watchpoints Worked examples mmult Thread affinity vectorization

X

XK6