ACM SIGACT News Distributed Computing Column 29 Idit Keidar Dept. of Electrical Engineering, Technion Haifa, 32000, Israel idish@ee.technion.ac.il Concurrent computing, once an esoteric pastime of Distributed Computing Theorists and High Performance Computing Extremists, is suddenly important in the computing world at large. This new interest in concurrency stems from a dramatic paradigm shift in computer architecture: No longer are hardware manufacturers making faster and faster (uni-)processors. Nowadays, chip companies are producing multi-processors with more and more cores. Only concurrent (multithreaded) programs can effectively exploit the potential of such multi-core processors. And thus, as multi-processors become mainstream, so does concurrent computing. This column features three contributions that re ect on the role of Distributed Computing research in the brave new world of multi-processors. What new challenges are raised by the newfound ubiquitous relevance of concurrency? What can we, the Distributed Computing community, do to address them? The most notable challenge stemming from the shift to multi-core architectures is the dif culty of programming concurrent code. One concept that tackles this challenge is transactional memory, which allows multiple processes (or threads) to concurrently access memory objects with transaction-like consistency semantics. As this concept is already taken seriously by industry, and
/lp/association-for-computing-machinery/acm-sigact-news-distributed-computing-column-29-2f6wnhzDQG