User Friendly High Productivity Computational Workflows Using the VISION/HPC Prototype

The complexity of parallel computing is daunting for new users and a productivity bottleneck for even experienced users. Furthermore, most users work from a Windows PC environment so that the overhead involved in learning Unix well enough to accomplish parallel programming tasks is itself a further obstruction to efficient productivity. What is needed is a smooth workflow that simplifies both the programming task and the remote execution management. VISION is a freely available, python-based, drag-and- drop visual-programming environment that reduces sophisticated programming tasks to drawing interactive flowcharts encapsulating the underlying programming complexity. Such flowcharts are built by dropping and connecting computational boxes on a flowchart instead of writing code. This is an important issue for productivity since productivity is dominated by the workflow of the user (e.g., time spent running versus time spent programming). As a python-based package, it is possible to embed parallel computing features from iPython into VISION to enable both visual programming as well as parallel execution on remote high performance computers. This paper discusses the prototype we built at SPAWAR Systems Center, San Diego for a visual parallel programming workflow based on VISION and iPython, for parallel computing using a Linux cluster as a backend, and a Windows XP workstation as the front-end running VISION.