*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Atlanta, GA | Posted: April 28, 2017
Albert "Ben" Snedeker, Communications Manager
Summary Sentence:
A School of CS student team recently earned a best student paper award at a prominent computer systems conference.
Full Summary:
No summary paragraph submitted.
Student best paper award EuroSys17Recognition of Georgia Tech in the fast-paced world of computer systems is continuing to grow.
This week in Belgrade, Serbia a team from the School of Computer Science earned the best student paper award at the 12th European Conference on Computer Systems (EuroSys17).
Presented by the conference’s program committee, the award recognizes work done by Ph.D. student Steffen Maass, postdoctoral fellow Changwoo Min, Ph.D. student Sanidhya Kashyap, postdoctoral fellow Woonhak Kang, Ph.D. student Mohan Kumar, and Assistant Professor Taesoo Kim.
Their paper titled, Mosaic: Processing a Trillion-Edge Graph on a Single Machine, details a new system for processing large-scale graphs that is demonstrated to be consistently faster than high-end distributed engines running across multiple computer clusters.
To achieve these results, the team coupled a novel approach to encoding graphs with fast storage media such as non-volatile memory express solid state drives (NVMe SSD) and massively parallel processors on a single machine. The team’s encoding method uses a new locality-optimizing, space-efficient graph representation that can be scaled up to meet larger computational needs.
“We envision this system to be a stepping stone toward systems which help to improve the performance and feasibility of large-scale processing for domains relying on graph processing, like machine learning,” said Maass.
“This should lead to researchers being able to run large-scale analysis on single machines rather than relying on more costly and complex cluster setups.”
A highlight of the team’s findings is that Mosaic can complete an iteration of Google’s PageRank algorithm on a trillion-edge graph in just 21 minutes. This is more than nine times faster than most distributed disk-based engines.
