VTune¶
Use Intel VTune Profiler to profile serial and multithreaded applications that are executed on a variety of hardware platforms (CPU, GPU, FPGA). The tool is delivered as a Performance Profiler with Intel Performance Snapshots and supports local and remote target analysis on the Windows, Linux, and Android* platforms. Without the right data, you’re guessing about how to improve software performance and are unlikely to make the most effective improvements. Intel® VTune™ Profiler collects key profiling data and presents it with a powerful interface that simplifies its analysis and interpretation.
Environmental models for VTune on ULHPC¶
bash
module purge
module load swenv/default-env/v1.2-20191021-production
module load toolchain/intel/2019a
module load tools/VTune/2019_update4
module load vis/GTK+/3.24.8-GCCcore-8.2.0
Interactive Mode¶
```bash
Compilation¶
$ icc -qopenmp example.c
Code execution¶
$ export OMP_NUM_THREADS=16 $ amplxe-cl -collect hotspots -r my_result ./a.out ```
To see the result in GUI use the command:
bash
amplxe-gui my_result
The command
bash
amplxe-cl
will list out the analysis types and $ amplxe-cl -hlep report will list out available reports in VTune.
Batch Mode¶
Shared Memory Programming Model (OpenMP)¶
```bash
!/bin/bash -l¶
SBATCH -J VTune¶
SBATCH -A ¶
SBATCH -N 1¶
SBATCH -c 28¶
SBATCH --time=00:10:00¶
SBATCH -p batch¶
module purge module load swenv/default-env/v1.2-20191021-production module load toolchain/intel/2019a module load tools/VTune/2019_update4 module load vis/GTK+/3.24.8-GCCcore-8.2.0
export OMP_NUM_THREADS=16 amplxe-cl -collect hotspots-r my_result ./a.out ```
Distributed Memory Programming Model¶
To compile just MPI application run $ mpiicc example.c and for MPI+OpenMP run $ mpiicc -qopenmp example.c.
```bash
!/bin/bash -l¶
SBATCH -J VTune¶
SBATCH -A ¶
SBATCH -N 2¶
SBATCH --ntasks-per-node=28¶
SBATCH --time=00:10:00¶
SBATCH -p batch¶
module purge module load swenv/default-env/v1.2-20191021-production module load toolchain/intel/2019a module load tools/VTune/2019_update4 module load vis/GTK+/3.24.8-GCCcore-8.2.0
srun -n ${SLURM_NTASKS} amplxe-cl -collect uarch-exploration -r vtune_mpi -- ./a.out ```
```bash
Report collection¶
$ amplxe-cl -report uarch-exploration -report-output output -r vtune_mpi
Result visualization¶
$ amplxe-gui vtune_mpi ``` The below figure shows the hybrid(MPI+OpenMP) programming analysis results:
Tip
If you find some issues with the instructions above, please report it to us using support ticket.