Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about OpenXLA's approach to achieving portable performance across diverse hardware architectures in this 17-minute conference talk. Explore how OpenXLA addresses the challenges of the heterogeneous hardware era, where different computing platforms require optimized solutions for machine learning workloads. Discover the technical strategies and implementation details that enable consistent performance across various hardware targets, from traditional CPUs and GPUs to specialized accelerators. Gain insights into the engineering principles behind portable optimization and understand how OpenXLA's compiler infrastructure adapts to different hardware characteristics while maintaining performance efficiency. Examine real-world applications and use cases that demonstrate the practical benefits of this portable performance approach in modern computing environments.
