[hpc-announce] Call for Abstracts -- Workshop on Modeling and Simulation of Systems and Applications ModSim 2019

[Hoisie, Adolfy]

Call for Abstracts

ModSim 2019
Workshop on Modeling & Simulation of Systems and Applications
August 14-16, 2019, University of Washington Botanic Gardens
Center for Urban Horticulture, Seattle

To promote advancements in modeling and simulation (ModSim) research, we are soliciting community input in the form of abstracts. If accepted, author(s) will be invited to offer a presentation at the annual gathering of our community, the ModSim 2019 Workshop.

Workshop URL: https://www.bnl.gov/modsim2019/ 
Submission URL: https://easychair.org/conferences/?conf=modsim20190 
EasyChair Submission Deadline: Friday, May 3, 2019 (11:59 PM, Anywhere on Earth [AOE])
Notification of Acceptance: Friday, May 31, 2019

Abstract Submission Guidelines
There is no set word limit for abstract submissions. However, please limit your submission to one page. The abstract should provide an overview that adequately summarizes the topic(s) presented and any proposed impact on ModSim research or techniques, especially related to full-system modeling and simulation. The following details a proposed abstract layout and points to consider:

Abstract Title

Primary research area:
-Modeling and Simulation of Artificial Intelligence and Machine Learning as a Method of ModSim
-Modeling and Simulation of Subsystems
-Integrative Methodologies and Tools for Full-system Modeling and Simulation

What is being modeled? (e.g., performance, reliability, power, other)
What is the target application?
What modeling techniques are being used?
What is novel about the approach versus current state of the art?
Are preliminary results available?

All abstracts must be submitted through EasyChair no later than Friday May 3, 2019 (11:59 PM, AOE). Those with accepted abstracts will be notified via e-mail on Friday, May 31, 2018.

ModSim 2019 Topics

The overarching theme for this year’s workshop is Full-system Modeling and Simulation. The emphasis will be on methodologies, tools, best practices, projects, and initiatives that aim to address the challenges and achieve the goal of modeling performance of full systems under a realistic application workload. Abstract contributions should focus on the following topical areas while maintaining a global view of “system” in consideration:

Modeling and Simulation of Artificial Intelligence and Machine Learning as a Method of ModSim

Artificial Intelligence (AI), in general, and Machine Learning (ML), in particular, have emerged as important application drivers in all forms of computing, including large-scale data- and numerically-intensive high-performance computing (HPC). The trend’s impact extends beyond the nature of architectures optimized for executing an ML workload. It also points the way toward applying AI/ML techniques as ModSim methodologies to support a range of systems (including but not limited to AI-centric systems). Thus, abstract submissions in this category should cover topics such as novel architectures to efficiently support AI/ML workloads, ML as a methodology for ModSim, and intelligent computational steering driven by dynamic and offline learning.

Integrative Methodologies and Tools for Full-system Modeling and Simulation

This category covers modeling and simulation technologies to predict performance, energy consumption, and cost of the whole system. A typical system comprises several subsystems with a breadth of architectures and exercises a broad spectrum of applications and workloads. It is highly desirable to optimize the whole system using ModSim with a level of effort commensurate with the resulting accuracy and precision. This workshop seeks abstracts that highlight how to further the state of the art, as well as future research directions in this field.

Modeling and Simulation of Subsystems

The blending of compute, memory devices, storage, and interconnect combined with application software profoundly impacts current and future computing systems in terms of performance, reliability, predictability, power consumption, and cost. Existing technologies are perceived as limited in terms of compute power, capacity, and bandwidth. Emerging technology approaches offer the potential to overcome both technology- and design-related limitations, addressing system requirements for many different applications. Modeling and simulating these subsystems is tremendously important, principally by affording the ability to characterize and quantify data movement, as well as estimate power consumption and other related behaviors in large-scale systems. Abstract submissions should relate to computer subsystem technologies and their characterization and/or provide use cases that describe how ModSim can help to overcome these significant challenges.