Adaptive Ensemble Biomolecular Simulations at Scale

04/12/2018
by   Vivek Balasubramanian, et al.
0

Recent advances in both theory and methods have created opportunities to simulate biomolecular processes more efficiently using adaptive ensemble simulations. Ensemble-based simulations are used widely to compute a number of individual simulation trajectories and analyze statistics across them. Adaptive ensemble simulations offer a further level of sophistication and flexibility by enabling high-level algorithms to control simulations-based on intermediate results. Novel high-level algorithms require sophisticated approaches to utilize the intermediate data during runtime. Thus, there is a need for scalable software systems to support adaptive ensemble-based applications. We describe three different types of adaptations and propose four operations necessary to support these adaptations. We identify and address the challenges of expressibility, instantiation and implementation of the four operations. We enhance the Ensemble Toolkit -- an ensemble execution system -- with abstractions to support these operations. We characterize the overhead of supporting adaptive capabilties in the toolkit at production scale. We implement two high-level adaptive algorithms: expanded ensemble and Markov state modeling. We highlight scientific advantages enabled by the novel capabilities of our approach and implementation on three distinct platforms.

READ FULL TEXT
research
04/12/2018

Implementing Adaptive Ensemble Biomolecular Applications at Scale

Many scientific problems require multiple distinct computational tasks t...
research
10/23/2017

Harnessing the Power of Many: Extensible Toolkit for Scalable Ensemble Applications

Many scientific problems require multiple distinct computational tasks t...
research
08/19/2022

Co-scheduling Ensembles of In Situ Workflows

Molecular dynamics (MD) simulations are widely used to study large-scale...
research
04/10/2021

Achieving 100X faster simulations of complex biological phenomena by coupling ML to HPC ensembles

The use of ML methods to dynamically steer ensemble-based simulations pr...
research
08/08/2019

From Piz Daint to the Stars: Simulation of Stellar Mergers using High-Level Abstractions

We study the simulation of stellar mergers, which requires complex simul...
research
01/18/2019

On-line Application Autotuning Exploiting Ensemble Models

Application autotuning is a promising path investigated in literature to...
research
06/01/2017

A Concurrency-Agnostic Protocol for Multi-Paradigm Concurrent Debugging Tools

Today's complex software systems combine high-level concurrency models. ...

Please sign up or login with your details

Forgot password? Click here to reset