{Open} MPI, Parallel Computing, the Universe, and Everything
Dr. Jeffrey M. Squyres
Thursday, November 7th, 2013 2:30pm
Lawrence Berkeley National Laboratory, Building 54 (Dining Hall)
Room 130 (Perseverance Hall)
Click here to RSVP
About the talk
Open MPI is a high performance, open source implementation of the MPI specification (BSD license FTW!). It is developed by a consortium of academic, research, and industry partners. This talk will discuss the current status of Open MPI, including some of its more advanced features, such as the new MPI-3 “MPI-T” control and performance variables, and the new, highly-flexible, user-specified process affinity system. This talk will also briefly discuss the ongoing work of expanding and evolving MPI by the MPI Forum (the standards body that controls the MPI specification).
Stay after the feature talk for a deep dive into Cisco’s new ultra low latency MPI transport for Ethernet starting at 4pm. Topics will include how the Cisco “usNIC” (userspace NIC) device uses SR-IOV to effect OS-bypass, how (for better or for worse) it is exposed to userspace via the Linux Verbs API, and some novel issues that arose while integrating the usNIC device into Open MPI.
About the speaker
Dr. Jeff Squyres is Cisco’s representative to the MPI Forum standards body and is Cisco’s core software developer in the open source Open MPI project. He has worked in the High Performance Computing (HPC) field since his early graduate-student days in the mid-1990’s, and is a chapter author of the MPI-2 and MPI-3 standards. Jeff received both a BS in Computer Engineering and a BA in English Literature from the University of Notre Dame in 1994; he received a MS in Computer Science and Engineering from Notre Dame two years later in 1996. After some active duty tours in the military, Jeff received his Ph.D. in Computer Science and Engineering from Notre Dame in 2004. Jeff then worked as a Post-Doctoral research associate at Indiana University, until he joined Cisco in 2006.
in the bottom right or navigate to