LLNL/damselfly

Name: damselfly

Owner: Lawrence Livermore National Laboratory

Description: Damselfly Network Simulator

Created: 2015-11-19 04:37:55.0

Updated: 2016-10-05 12:22:50.0

Pushed: 2016-03-09 03:53:26.0

Homepage: null

Size: 1845

Language: Objective-C

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

Damselfly v1.0

Damselfly is a model-based parallel network simulator. It can simulate communication patterns of High Performance Computing applications on different network topologies. It outputs steady-state network traffic for a communication pattern, which can help in studying network congestion and its impact on performance.

Limitation: can be run on up to 16,384 MPI processes. The primes.h file has a limited number of entries.

Build
r build && cd build
e ..

Sample cmake command for Blue Gene/Q:

e -DCMAKE_C_COMPILER=mpixlc_r -DCMAKE_CXX_COMPILER=mpixlcxx_r -DCMAKE_BUILD_TYPE=Release -DJOB_TRAFFIC=ON ..
Run
un -np 8 ./damselfly conffile map_file out_file

where:

Format of conffile
al number of routers
ups #rows #columns #nodes #cores
name for intra group connection (binary format)
name for inter group connection (binary format)
name #ranks #edges (binary)
name #ranks #edges (binary)
as many files as needed

Code for generating sample connection files are in the utils directory.

Reference

Any published work that utilizes this software should include the following reference:

il Jain, Abhinav Bhatele, Xiang Ni, Nicholas J. Wright, and Laxmikant V.
. Maximizing Throughput on a Dragonfly Network. In Proceedings of the
rnational Conference for High Performance Computing, Networking, Storage
Analysis, SC '14, November 2014. LLNL-CONF-653557.
Release

Copyright (c) 2014, Lawrence Livermore National Security, LLC. Produced at the Lawrence Livermore National Laboratory.

Written by:

Nikhil Jain <nikhil.jain@acm.org>
Abhinav Bhatele <bhatele@llnl.gov>
Peer-Timo Bremer <ptbremer@llnl.gov>

LLNL-CODE-678961. All rights reserved.

This file is part of Damselfly. For details, see: https://github.com/LLNL/damselfly. Please also read the LICENSE file for our notice and the LGPL.


This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.