ceClub: OS Services for Computational Accelerators

Mark Silberstein (EE, Technion)
Wednesday, 15.1.2014, 11:30
Room 337-8 Taub Bld.

Future applications will need to use computational accelerators like GPUs to achieve their performance and power goals. However building efficient systems that use accelerators today is incredibly difficult. The main problem lies in the lack of appropriate OS support -- while OSes provide optimized resource management and Input/Output (I/O) services to CPU applications, they make no such services available to accelerator programs.

In this talk I will discuss my ongoing work on building an Operating System layer for GPUs that enables access to files and network services directly from programs running on GPUs. I will describe the existing infrastructure, my ongoing work, open questions and future directions.

The talk is self-contained, no background in GPU computing is necessary. Bio:
Mark Silberstein is an Assistant Professor at EE, Technion. He truly believes that building practical, programmable and efficient computer systems with computational accelerators requires cross-cutting changes in system interfaces, OS design, hardware security mechanisms, storage and networking services, as well as programming models and parallel algorithms. He is eagerly looking for excellent graduate students to join his efforts on emancipation of accelerators in computer systems. Web page: https://sites.google.com/site/silbersteinmark

Back to the index of events