As conventional von–Neumann architectures are suffering from rising power densities, we are facing an era with power, energy efficiency, and cooling as first–class constraints for scalable HPC. FPGAs can tailor the hardware to the application, avoiding overheads and achieving higher hardware efficiency than general–purpose architectures. Leading FPGA manufacturers have recently made a concerted effort to provide a range of higher–level, easier to use high–level programming models for FPGAs. Such initiatives are already stimulating new interest within the HPC community around the potential advantages of FPGAs over other architectures. With this in mind, this workshop, now in its third year, brings together HPC and heterogeneous–computing researchers to demonstrate and share experiences on legacy and new high–level programming models, optimizations specific to scientific computing and data analytics, tools for performance/energy improvements, FPGA computing in the cloud, and popular applications for reconfigurable computing such as machine learning and big data.<br>Submissions are solicited that explore the state of the art in the use of FPGAs in heterogeneous high–performance computing architectures and, at a system level, in data centers and supercomputers. FPGAs may be considered from either or both the distributed, parallel and composable fabric of compute elements or from their dynamic reconfigurability. We particularly encourage submissions which focus on the mapping of algorithms and applications to heterogeneous FPGA–based systems as well as the overall impact of such architectures on the compute capacity, cost, power efficiency, and overall computational capabilities of data centers and supercomputers. Submissions may report on theoretical or applied research, implementation case studies, benchmarks, standards, or any other area that promises to make a significant contribution to our understanding of heterogeneous high–performance reconfigurable computing and will help to shape future research and implementations in this domain. A non–comprehensive list of potential topics of interest is given below:<br>FPGAs in Supercomputer, Cloud and Data Center: FPGAs in relation to challenges to Cloud/Data Center/Supercomputing posed by the end of Dennard scaling<br>Supercomputing, Cloud and Data Center Applications: Exploiting FPGA compute fabric to implement critical cloud/HPC applications<br>Leveraging Reconfigurability: Using reconfigurability for new approaches to algorithms used in cloud/HPC applications<br>Benchmarks: Compute performance and/or power and cost efficiency for cloud/HPC with heterogeneous architectures using FPGAs<br>
Abbrevation
H2RC
City
Denver
Country
United States
Deadline Paper
Start Date
End Date
Abstract