December 7, 2021 – In the final episode of the Let’s Talk Exascale podcast, Scott Gibson interviews Harry Yoo, Thomas Brettin and Venkatram Vishwanath of the Argonne National Laboratory and the CANDLE Project. The podcast was released on December 7, 2021.
Hello and welcome to episode 91 of the Let’s talk about Exascale Podcast. It is here that we explore the efforts of the Department of Energy’s exascale computing project, from development challenges and achievements to the expected ultimate impact of exascale computing on society.
And this is the fourth in a series of work-based episodes aimed at sharing best practices in preparing applications for the upcoming Aurora exascale supercomputer at the Argonne Leadership Computing Facility.
The series highlights achievements in optimizing code to run on GPUs. We also provide lessons learned to developers to help them overcome initial hurdles.
This time, we are focusing on the computer codes used in a project called CANDLE, which stands for CANcer Distributed Learning Environment. It addresses three important scientific challenges in cancer research, and we will hear about them shortly. The focus of the work is on machine learning and in particular relies on a single scalable deep neural network code, or DNN, which also goes by the name CANDLE. The project is developing highly efficient DNNs optimized for the unique architectures provided by exascale class computing platforms such as future Aurora and Frontier systems.
The CANDLE project is a collaborative effort with the US Department of Energy and the National Cancer Institute (NCI), involving Argonne, Lawrence Livermore, Los Alamos, and Oak Ridge National Laboratories.
Guests of the program are Thomas Brettin, Venkatram Vishwanath and Harry Hyunseung Yoo from the Argonne National Laboratory and the CANDLE project.
Our subjects: an overview of the three project challenges, how CANDLE will benefit from exascale computer systems, the role of PCE in the development of CANDLE, and more.
Link to listen and access a full transcript:
Source: Scott Gibson, Exascale IT project