Saltar al contingut Menu
Map
  • Home
  • Information
  • Contact
  • Map

The Structural Simulation Toolkit: A Scalable Architecture Simulation Framework for Understanding both Exascale Algorithms and Exascale Systems

Compartir
Introduced: 15-03-2016
HPC (CAP) research group invites you to attend the talk:
Speaker: Jeremiah Wilke (Sandia National Lab)
Date: Fri, 8/Apr/2016, 09:30
Room: C6-E101
ABSTRACT

The Structural Simulation Toolkit (SST) serves Sandia as a tool to accelerate the design cycle for high performance computing (HPC) systems. Hardware trends have guided algorithms and new algorithms have guided hardware trends. Simulation allows speculative hardware and speculative algorithms to be tested together, instead of isolation, accelerating the design cycle. While co-design strategies have proven effective for embedded systems, system co-design based on simulation poses challenges for exascale systems. Simple analytic or constitutive models can be easily scaled and provide some insight. Structural models based on discrete event simulation provide much greater flexibility and accuracy. Structural simulations requires parallel discrete event simulation (PDES) for large scales, which can be difficult to implement efficiently. Coupling distinct simulation models from different code bases while enabling PDES scaling quickly becomes an intractable problem. SST enables scalable simulations incorporating flexible and modular simulation components through a unified PDES core. We broadly discuss the growing SST ecosystem and some recent successes. We further discuss the particular challenge of simulating dynamic, asynchronous runtime systems at extreme scales. We finally present the SST macroscale components for developing exascale software stacks supporting asynchronous runtimes.

Conference Information


Compartir

 
logo FIB © Barcelona school of informatics - Contact - RSS
This website uses cookies to offer you the best experience and service. If you continue browsing, it is understood that you accept our cookies policy.
Classic version Mobile version