The Texas Advanced Computing Center (TACC) at The University of Texas at Austin is seeking applications from research teams wishing to participate in the Early Science Program for the "Knight's Landing" (KNL) upgrade of Stampede to the second generation of the Intel® Xeon Phi™ processor.
The Stampede supercomputer at TACC is one of the leading systems in the National Science Foundation (NSF) open cyberinfrastructure, and was the first system to make the original Xeon Phi™ coprocessor available to the research community at large scale. A portion of the Stampede system will be upgraded to the second generation of Xeon Phi™, codenamed "Knight's Landing," in the first half of 2016 in preparation for future systems.
TACC is seeking a set of research teams that would like early access to this technology, and are willing to work closely with TACC in preparing their codes to take full advantage of the next Xeon Phi™. TACC staff members will assist successful teams to profile and analyze their application on existing Xeon and Xeon Phi™ platforms, developing and executing a plan to enhance the code for the KNL processor. Selected teams will receive early access and significant compute time on the KNL nodes on Stampede before they are put into production for general access, as well as a commitment of time from TACC technical staff. Selected teams will be expected to collaborate with TACC on code improvements and benchmarking.
The Stampede Early Science Program will cooperate with similar early adopter programs for Department of Energy Xeon Phi-based systems at Argonne National Lab and the National Energy Research Science Center at Lawrence Berkeley National Labs. Care will be taken in the selection of proposals to avoid duplication of efforts between these programs.
Projects will begin on a rolling start basis in January 2016. Approximately six to eight teams will be supported. Requests to join the Early Science Program can be made by sending an email to firstname.lastname@example.org. A request should include a brief description of the application to be optimized, including current scaling data, and a brief description of the science challenges to be addressed.
For further information, please contact email@example.com.