Search

We describe the design, deployment and operation of a computer system built to efficiently run deep learning frameworks. The system consists of 16 IBM POWER9 servers with 4 NVIDIA V100 GPUs each, interconnected with Mellanox EDR InfiniBand fabric, and a DDN all-flash storage array. The system is tailored towards efficient execution of the IBM Watson Machine Learning enterprise software stack that combines popular open-source deep learning frameworks. We build a custom management software stack to enable an efficient use of the system by a diverse community of users and provide guides and recipes for running deep learning workloads at scale utilizing all available GPUs. We demonstrate scaling of a PyTorch and TensorFlow based deep neural networks to produce state-of-the-art performance results.

The first programming language for most children worldwide today is a blocks-based language like Scratch or Snap. GP is a blocks-based programming language designed to be a follow-on. GP aims to be a general purpose language, like Python or Ruby, in which students can create more complex programs or create standalone applications. GP includes a mechanism for creating modules for others to re-use and a wide variety of primitives, including tools for manipulating CVS and JSON data, using serial port and network connections, and manipulating pixels in pictures or samples in sounds. The language could be used in computer science classes beyond introductory computer science. GP explores the questions: "How far can we go with a blocks-based programming language? Do we have to move students to a textual programming language to explore advanced computational ideas and applications?" In this laptop required workshop, participants will try out GP. They will explore sample projects and create their own projects that push on advanced features of GP such as using multiple classes and instances, creating sets of blocks that can be shared as extension modules, using cloud data, or manipulating images and sounds. GP will be released (free and open source) in Summer 2017, so our goal is to find early adopters who are interested in trying GP and developing examples for others.

This article distills a discussion about the goals, mechanisms, and effects of three environments which aim to support the acquisition and development of computing concepts (problem solving and programming) in pre-University and non-technical students: Alice, Greenfoot, and Scratch. The conversation started in a special session on the topic at the 2010 ACM SIGCSE Symposium on Computer Science Education and continued during the creation of the resulting Special Issue of the ACM Transactions on Computing Education.

This panel will showcase and compare three leading Initial Learning Environments (ILE): Alice, Greenfoot and Scratch.