Implementing Omni OpenMP Runtime with Argobots

Event Sponsor: 
Mathematics and Computer Science Division
Start Date: 
Feb 5 2016 - 10:30am
Building 240/Room 4301
Argonne National Laboratory
Jinpil Lee
Speaker(s) Title: 
RIKEN, Japan
Sangmin Seo

The increasing number of cores in a chip and complex memory hierarchy requires multithreading programming models to handle nested parallelism and dynamic task management. In this presentation, we'll introduce our Omni OpenMP compiler, which is currently being re-implemented using Argobots. Some benchmark evaluation data will be provided to show the performance of Argobots API and OpenMP constructs which are implemented on it.

Jinpil Lee received his PhD degree in computer science from University of Tsukuba in 2013, under the supervision of Prof. Mitsuhisa Sato. From 2013 to 2015, he was working in KISTI, the national supercomputing center in Korea. Currently he is working at Riken AICS in Japan, doing research about directive-based parallel programming models.