Lesson 1: Learn about GPUs and efficient data processing


H2O is an open source predictive analytics platform for data scientists and business analysts who need scalable and fast machine learning. Unlike traditional analytics tools, H2O provides a combination of extraordinary math and high performance parallel processing with unrivaled ease of use. H2O speaks the language of data science with support for R, Python, Scala, Java and a robust REST API. Smart business applications are powered by H2O’s nano-fast scoring engine.

High processing efficiency demands greater hardware capabilities and the conventional ways of increasing speed was to employ distributed computing on multi-core CPUs with faster clock rates. The exponential increase in data available for processing and complex architectures out-ran the processing capabilities of a CPU and made way for GPUs which were inherently thought to just be good for processing graphics. The industry was quick to identify the fine-grained parallelism in GPUs’ architecture and utilized it for general purpose computing. GPUs bring in a 10x increase in computing performance and a 5x increase in energy efficiency as the architecture is tolerant of memory latency and has a larger number of transistors dedicated to computation.

The following webinar recording by Wen Phan, Senior Solutions Architect at H2O.ai, discusses the associated enabling technologies, such as CUDA, and demonstrates GPU-expedited performance with the H2O platform, Deep Water.





Questions?

  • We'd love to hear from you! Reach out to us on our Google Stream groupwith questions, requests, or comments any time.