Entrambe le parti precedenti la revisione
Revisione precedente
Prossima revisione
|
Revisione precedente
|
calcoloscientifico:vtune [05/05/2018 09:44] paolo.oggiano [VTune] |
calcoloscientifico:vtune [11/05/2018 10:34] (versione attuale) paolo.oggiano [HPC (UNIPR)] |
==== Collect ==== | ==== Collect ==== |
| |
Per la raccolta dei dati di profiling occorre lanciare l'eseguibile tramite amplxe-cl, | To collect profiling data, run the executable via amplxe-cl, |
specificando una action tra le seguenti: https://software.intel.com/en-us/vtune-amplifier-help-amplxe-cl-actions | specifying an action between the following: https://software.intel.com/en-us/vtune-amplifier-help-amplxe-cl-actions |
| |
L'action necessaria per raccogliere dati riguardo le performance e' **collect**. | The action necessary to collect data about performance is **collect**. |
Questa action richiede una opzione per specificare il tipo di analisi, tra cui: | This action requires an option to specify the type of analysis, including: |
* **hotspots**: Identify your most time-consuming source code. | * **hotspots**: Identify your most time-consuming source code. |
* **concurrency**: Analyze how your application is using available logical CPU cores, discover where parallelism is incurring synchronization overhead, and identify potential candidates for parallelization. | * **concurrency**: Analyze how your application is using available logical CPU cores, discover where parallelism is incurring synchronization overhead, and identify potential candidates for parallelization. |
* **memory-access**: Measure a set of metrics to identify memory access related issues (for example, specific for NUMA architectures). | * **memory-access**: Measure a set of metrics to identify memory access related issues (for example, specific for NUMA architectures). |
| |
Esempio di comando: | Command example: |
amplxe-cl -collect hotspots -result-dir <dirname> home/test/myApplication | amplxe-cl -collect hotspots -result-dir <dirname> home/test/myApplication |
| |
=== Analisi dei risultati === | === Result analysis === |
| |
L'analisi off-line dei dati raccolti puo' essere fatta in modo grafico con amplxe-gui: | The off-line analysis of the collected data can be done in a graphical way with amplxe-gui: |
| |
amplxe-gui <dirname> | amplxe-gui <dirname> |
==== HPC (UNIPR) ==== | ==== HPC (UNIPR) ==== |
| |
Connettersi a gui.hpc.unipr.it ed eseguire un job interattivo; ad esempio: | Connect to gui.hpc.unipr.it and perform an interactive job; eg: |
qsub -I -l select=1:cluster=knl:ncpus=68 | srun -N1 -n68 -p knl --account=<account name> |
| |
> module load intel # (include anche vtune) | > module load intel # (include vtune too) |
> amplxe-cl -collect hotspots -result-dir <dirname> home/test/myApplication | > amplxe-cl -collect hotspots -result-dir <dirname> home/test/myApplication |
> exit | > exit |
| |
Sul cluster HPC di Ateneo e' preferibile lanciare amplxe-gui da gui.hpc.unipr.it: | It is preferable to launch on the University HPC cluster amplxe-gui from gui.hpc.unipr.it: |
| |
module load intel | module load intel |
=== MARCONI (CINECA) === | === MARCONI (CINECA) === |
| |
Su Marconi si puo' sottomettere una sessione interattiva, in cui eseguire sia la collect che la data analysis (amplxe-gui non funziona sulla macchina di login), come ad esempio: | On Marconi you can 'submit an interactive session, in which to perform both the collect and the data analysis (amplxe-gui does not work on the login machine), such as: |
srun -N1 -n68 -t 00:30:00 -A <account> -p knl_usr_dbg --pty bash | srun -N1 -n68 -t 00:30:00 -A <account name> -p knl --pty bash |
| |
> module load intel vtune | > module load intel vtune |
=== Debug Info === | === Debug Info === |
| |
Per includere informazioni di debug nei risultati (ad esempio il codice sorgente) occorre compilare con l'opzione | To include debugging information in the results (for example the source code) you need to fill in with the option |
| |
<code> | <code> |