DL Framework Developer Toolkit | Intel® Tiber™ Developer Cloud

Before you begin, please consider reading the Get Started with the Intel® oneAPI Base Toolkit on the Intel® DevCloud document if you haven’t already done so.

Deep learning frameworks offer optimized building blocks to streamline designing, training, and validating deep neural networks through a high-level programming interface.

Build or customize deep learning frameworks using optimized deep learning libraries
Optimized for high performance on Intel CPUs and GPUs
Target single-node or multi-node distributed processing with common APIs that allow scaling across nodes

The following sample can be executed either using a JupyterLab* session or the SSH terminal. For more information on how to connect to the Intel DevCloud using either Jupyter or SSH please visit https://devcloud.intel.com/oneapi/get_started/.

Jupyter

Connect with JupyterLab
Open a terminal.

Download the samples.

git clone https://github.com/oneapi-src/oneAPI-samples.git

Go to the sample location.

cd ~/oneAPI-samples/Libraries/oneDNN/tutorials

Open the Jupyter based get started guide and follow along.
```
tutorial_getting_started.ipynb
```

SSH Terminal

Download the Intel® oneAPI Base Toolkit Samples

Connect to the DevCloud.
```
ssh devcloud
```

Download the samples.

git clone https://github.com/oneapi-src/oneAPI-samples.git

Create the job scripts

Go to the sample location.

cd ~/oneAPI-samples/Libraries/oneDNN/getting_started

Create a build.sh script with the following contents.

#!/bin/bash
source /opt/intel/inteloneapi/setvars.sh > /dev/null 2>&1
mkdir dpcpp
cd dpcpp
cmake ..
make

Create a run.sh script with the following contents for executing the sample.

#!/bin/bash
source /opt/intel/inteloneapi/setvars.sh > /dev/null 2>&1
./build/bin/getting-started-cpp cpu
./build/bin/getting-started-cpp gpu

Build

Build the sample on a gpu node.
```
qsub -l nodes=1:gpu:ppn=2 -d . build.sh
```

Run

Run the sample on a gpu node.
```
qsub -l nodes=1:gpu:ppn=2 -d . run.sh
```

Monitor jobs

In batch mode, the commands return immediately; however, the job itself may take longer to complete. In order to inspect the job progress, use the qstat utility.
```
watch -n 1 qstat -n -1
```
Note: The watch -n 1 command is used to run qstat -n -1 and display its results every second.

Check the results

Upon completion, the stderr and stdout of the job are written to the disk:
```
<script_name>.sh.eXXXX, which is the job stderr
<script_name>.sh.oXXXX, which is the job stdout
```
Here XXXX is the job ID, which gets printed to the screen after each qsub command.
Inspect the output of the sample.
```
cat run.sh.oXXXX
```
Remove the stdout and stderr files.
```
rm run.sh.*
```

Additional information

More information about the architecture of the Intel® DevCloud and about the Portable Batch System can be found at the following locations:

Intel® Tiber™ Developer Cloud for oneAPI