| Auszug |
|---|
| Codeblock |
|---|
|
| vtune -report summary -r vtune_* Intel VTune, APS, Intel Advisor, Intel Inspector, Intel Trace Analyzer and Collector
("vtune", "advisor", "inspector", and "itac" environment modules) |
VTune
Intro:
https://www.intel.com/content/www/us/en/docs/vtune-profiler/get-started-guide/2023/linux-os.html
Help:
Run VTune via command line interface (recommended non-interactive use)
Run your application with VTune wrapper as follows:
www.intel.com/content/www/us/en/docs/vtune-profiler/user-guide/2023-0/command-line-interface.html
Example
| Codeblock |
|---|
|
mpirun -np 4 vtune –collect hotspots -result-dir vtune_hotspot ./path-to_your/app.exe args_of_your_app |
After completion explore hotspot analysis e.g. via
| Codeblock |
|---|
|
vtune -report summary -r vtune_* |
VTune-GUI (not recommended)
Login with x-window support (ssh -X) and then start
Run VTune-GUI remotely on your own browser (recommended interactive use)
for x86 based hardware: best for Intel, some AMD) |
| module load | use case |
|---|
VTune (← click for more)
- Application Performance Snapshot
- Profiler
| vtune/ |
- high level - initial overview
- low level - detailed performance analysis (e.g. hotspots, bottlenecks)
|
| Advisor | advisor/ | low level - threading and vectorization aid (e.g roofline analysis) |
| Inspector | inspector/ | low level - memory & threading error checking (e.g. find leaks or data races) |
| Trace Analyzer and Collector | itac/ | low level - identify MPI load imbalances and communication hotspots |
Documentation is provided here: Materials