By projecting the increase in performance of the world’s top supercomputers (measured in floating-point operations per second, or FLOPS) the exa-scale level (1018 FLOPS) is likely to be reached by 2020. However, the goal is to obtain exa-FLOPS performance staying within a 20 Mega-Watt power envelope. This leads to a performance efficiency requirement of 50 GFLOPS/W, which is about 20 times higher than where we are today.
Moreover, a number of other issues need to be addressed to meet the so called “exascale challenge”. These include: processor architecture and extreme parallelism, memory bandwidth and capacity, programming models for high-concurrency, harder to manage I/O system, reliability and resiliency.
We invite researchers to submit contributions in current research topics which projected in the future will meet the exascale challenge.
The workshop’s focus is on the following areas:
• fault-tolerant and self-healing computing
• testing of large and complex systems
• dependability
• power and thermal management of multicore chips
• energy efficient computing
• emerging technologies
Groups and researchers are invited to submit an abstract, maximum one page,
explaining how their research will contribute to the exascale challenge.
Submissions should be sent as plain text, or as PDF attachments, by
e-mail to:
median-exascale@compute.dtu.dk
Authors of accepted contributions will be invited to present their ideas
at the workshop in the form of a 20/25 minute oral presentation.
Based on the availability of MEDIAN funding, a number of travel reimbursement
grants will be available and prioritized as follows:
1. Authors of accepted contributions
2. MEDIAN members
3. COST countries balancing
4. First-come, first-served basis
IMPORTANT DATES
Submission Deadline October 25, 2013
Acceptance notification November 1, 2013
Registration deadline November 27, 2013
WORKSHOP ORGANIZATION
Alberto Nannarelli, Technical University of Denmark
Call for paper: pdf