Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the theoretical and practical challenges of designing mixed precision algorithms in numerical linear algebra through this 27-minute conference talk by Erin C. Carson from Charles University, Czech Republic. Understand how mixed precision arithmetic is becoming a key feature of modern high-performance computing architectures and learn strategies for safely and effectively leveraging this trend in computational science. Discover why mixed precision matters for future hardware development and examine how low precision formats can be utilized without sacrificing computational accuracy, particularly when other sources of inexactness like discretization or measurement error are already present in the system. Analyze methods for balancing error sources in scientific computing applications and review real-world case studies demonstrating how mixed precision approaches can significantly boost computational performance while maintaining reliability.
