-
Notifications
You must be signed in to change notification settings - Fork 16
home
Discrete event driven simulation of HPC system architectures and subsystems has emerged as a productive and cost-effective means to evaluating potential HPC designs, along with capabilities for executing simulations of extreme scale systems. The goal of the CODES project is use highly parallel simulation to explore the design of exascale storage/network architectures and distributed data-intensive science facilities.
Our simulations build upon the Rensselaer Optimistic Simulation System (ROSS), a discrete event simulation framework that allows simulations to be run in parallel, decreasing the simulation run time of massive simulations to hours. We are using ROSS to explore topics including large-scale storage systems, I/O workloads, HPC network fabrics, distributed science systems, and data-intensive computation environments.
Historically, the CODES project has been a collaboration between the Mathematics and Computer Science department at Argonne National Laboratory and Rensselaer Polytechnic Institute. We have collaborated with researchers at University of California at Davis to come up with novel methods for analysis and visualizations of large-scale event driven simulations. We have also collaborated with Lawrence Livermore National Laboratory for modeling HPC interconnect systems.
- PDES and ROSS: Starting from Scratch
- CODES Overivew
- Installing CODES
- Synthetic Workload Development
- CODES Inner Workings: How does CODES actually work?
- CODES Development Ideology
- Execution Example: 1D Dragonfly
Instructions for current in-development branches.
Some of the links below might contain woefully outdated info. They are kept for completeness reasons: CODES is (or was once) capable of all things mentioned below; the precise instructions to run some of the mentioned capabilities are different now than what they were when the instructions were written.