Our worlds are full of massive amounts of complexly connected data. Graph analytics offer a solution for extracting meaning from the mass. Applied examples include: discovering malicious exfiltrators in networks with millions of connections, or finding optimal drug targets in elaborate chemical pathways.
Multiple low-level GPU-accelerated graph frameworks exist, but a steep learning curve has prevented widespread use of iterative graph analytics by the data science community. This session will focus on PyDASL, a Python tool capable of performing GPU accelerated graph operations. PyDASL enables developers to quickly author iterative graph analytics with up to 9,000% less code than other comparably performing GPU-accelerated graph frameworks. We will show the ease of writing and executing a new graph analytic and review PyDASL’s built-in analytics.