دانلود کتاب کتابخانه‌های نرم‌افزاری عددی و تأثیر آن‌ها بر کاربردهای علمی

Some of the most significant products of applied mathematics are algorithms that enable faster and more accurate science and engineering simulations, or even the solution of problems that were previously impossible to model. These algorithms most often impact simulations through numerical software libraries. Numerous numerical software libraries have high-quality implementations of efficient algorithms which facilitate transfer of new algorithms into scientific applications, including those in laboratory and industry settings. Recently, use of these libraries within science and engineering applications has significantly increased, showing the value of these libraries more than ever before. Particularly visible is the context of the recent U.S. Department of Energy’s Exascale Computing Project (ECP) where the use of libraries has enabled applications to migrate their simulation codes to take advantage of the newest exascale systems as well as improve performance and capabilities on more traditional architectures. During the 2023 International Congress on Industrial and Applied Mathematics, a three-part minisymposium was held on the topic of, “Numerical Software Libraries Enabling Benefits to Scientific Applications.” These sessions focused on numerical library activities undertaken to ensure value to applications and gave examples demonstrating the use of new algorithms as well as capability and speed improvements within many application areas. The sessions underscored the significant impact of applied mathematics on science and engineering applications. This volume opens with a chapter describing the extreme-scale software development kit (xSDK), a collection of mathematical libraries and a subset of a larger software ecosystem developed within the ECP. Many large-scale applications, such as multi-physics simulations, require the use of a variety of different independently developed libraries in combination. Ensuring seamless builds and reliable performance of whole applications which utilize more than one of the xSDK libraries requires a healthy software ecosystem. Chapter one discusses the importance of such an ecosystem, including community-driven guidelines to facilitate sustainability and interoperability and the positive impact of such an ecosystem on applications, with focus on the xSDK.