In this episode of the O’Reilly Podcast, I speak with Amit Vij, CEO and co-founder of Kinetica, a company that has developed an analytics database that uses graphics processing units (GPUs). We talk about how organizations are using GPU-accelerated databases to converge artificial intelligence (AI) and business intelligence (BI) on a single platform.
- The benefits of converging AI and BI in a single system: “You are orders of magnitude faster, and you have the ability to operate on real-time data, as opposed to operating on yesterday’s data,” Vij says.
- The processing speed of GPUs: “The GPU really leverages parallel processing so you can maximize your throughput and take advantage of the advancements in hardware that have come about,” he says.
- How GPU databases break down the walls between the data science and business domains: “Nowadays machine learning scientists and mathematicians can, in just three lines through SQL, execute their algorithms directly on data sets that are billions of objects,” Vij says.
- How GPU databases integrate with machine learning tools such as TensorFlow, and how GPU applications can use the cloud