With the rise of deep learning GPU-based computation is becoming more and more popular. However, GPU programming has a much broader applicability and we sometimes run into situations that are not covered by standard frameworks and libraries. Because parallel programming on GPUs can be an exotic topic for many people in our field, we wanted to prepare a proper introduction for our DataScience@UL-FRI community! On Thursday we held a workshop titled Parallel computation on GPUs with OpenCL. Rok Češnovar and Davor Sluga provided a great explanation of the art of GPU programming, supported by practical examples, ranging from simple problems, such as vector addition, to more advanced ones, such as image processing.