By Ulrike Meier Yang
The pursuit of exascale promises larger and faster computers capable of performing increasingly more complex simulations at unprecedented scale, leading to better science. Achieving this goal requires not only technical advancements, such as hardware with faster components and larger memories, but also advanced software capable of taking full advantage of the hardware and complex scientific models. It also requires highly skilled experts from many disciplines working collaboratively to deliver the necessary technical and scientific components. The Exascale Computing Project (ECP) provides such an environment for exascale research and development.
On a smaller scale, the Extreme-scale Scientific Software Development Kit (xSDK) project, one of ECP’s projects, works toward the seamless build and use of a variety of independently developed highly efficient interoperable math libraries to support scientific applications. As a member of the xSDK project, I have participated in its growth and collaborated with many different talented software developers, mathematicians, and computer scientists from a variety of institutions, located both inside and outside the United States.
Over the years, we have overcome communication barriers and created a community through regular, mostly virtual meetings. Having the skills to work and collaborate online has allowed us to continue our work successfully despite the COVID-19 pandemic. Being part of the ECP-facilitated access to other specialists, such as developers of runtime systems, performance tools, application developers, etc., has led to additional external collaborations that are crucial to achieving our project’s goal.
While technical advancements are essential to achieving exascale computing, the human factor is just as important. Working together toward a common goal, we share successes and new discoveries that help us refine our software. Helping improve each other’s libraries while giving credit where credit is due increases the quality of the whole product and moves us closer to the goal of achieving great science at exascale. Everybody wins—not only now but also in the future—because the relationships that have been formed during ECP will continue to thrive beyond the life of the project and benefit research endeavors for many years to come.