Accelerating GAN training using highly parallel hardware on public cloud
This work addresses the need for efficient and cost-effective infrastructure for machine learning in high-energy physics, though it is incremental as it applies existing parallelization methods to a specific domain.
The paper tackled the problem of accelerating GAN training for high-energy physics applications by parallelizing it on multiple GPUs and TPUs in public cloud environments, achieving linear speed-up while maintaining performance comparable to Monte Carlo simulations.
With the increasing number of Machine and Deep Learning applications in High Energy Physics, easy access to dedicated infrastructure represents a requirement for fast and efficient R&D. This work explores different types of cloud services to train a Generative Adversarial Network (GAN) in a parallel environment, using Tensorflow data parallel strategy. More specifically, we parallelize the training process on multiple GPUs and Google Tensor Processing Units (TPU) and we compare two algorithms: the TensorFlow built-in logic and a custom loop, optimised to have higher control of the elements assigned to each GPU worker or TPU core. The quality of the generated data is compared to Monte Carlo simulation. Linear speed-up of the training process is obtained, while retaining most of the performance in terms of physics results. Additionally, we benchmark the aforementioned approaches, at scale, over multiple GPU nodes, deploying the training process on different public cloud providers, seeking for overall efficiency and cost-effectiveness. The combination of data science, cloud deployment options and associated economics allows to burst out heterogeneously, exploring the full potential of cloud-based services.