Author: wp_738037
The goal of this lecture is to present and compare the features of some of the most popular Big Data programming frameworks for HPC systems, such as Hadoop, Spark, Storm, Airflow, MPI, Hive, and Pig,…
The lecture shows new trends on HPC architectures, which will be based on heterogenous platforms mixing GPPs (General-Purpose-Processor) and accelerators. At the GPP side, new multi-core Arm 64b solutions with scalable vector extension recently emerged…
Recent advances in hardware technology led to availability of more and more powerful architectures. On the single system scenario, processors with an impressive number of cores and/or hardware thread contexts are available, often equipped with…
We discuss the history of MPI and its alternatives, in light of recent hardware and software developments. We will discuss how RMA features of modern hardware allow for the evolution of Partitioned Global Address Space…