forsyde / DeSyDe

This repository contains our design space exploration tool.

DeSyDe

DeSyDe is a design space exploration tool developed at KTH (ForSyDe research group).

Releases:

• latest:

  • Release for our DSD'18 publication + user tutorial

• previous:

  • Release for our TODAES article
  • Release for our RAPIDO'17 publication

(Almost) hassle-free installation

You need to install DeSyDe via the automated build scripts. We have tried assuring an (almost) fully-automated installation process, especially for Linux machines. The idea is quite simple: the script downloads almost everything necessary so that nothing on your system is touched and then proceeds to compile everything. This sandboxing comes with the cost of added compilation time, but since this should be a one-time process, the larger time frame is a good trade-off for flexibility.

The only dependency that is not cloned directly from its repo and compiled alongside DeSyDe is Qt, as DeSyDe currently does not make Gecode's Gist optional. Please ensure that you have the basic development files for Qt installed and reachable in your machine. In future releases this necessity will be removed.

If you are on any debian-based distro with reasonably updated packages, you should be good to go by issuing the following install command (do not forget to prepend sudo if necessary):

apt install automake libtool qt5-default

As of 2019-06-05, it seems from user feedback that on ubuntu and other derived distros not all dependencies are pulled with these commands, so it may be necessary to install qtcreator to be able to compile DeSyDe (do not forget to prepend sudo if necessary):

apt install qtcreator

Then, a make followed by make install should do the trick. Tested on Linux Mint 18.3 and Debian 10.

Usage

Please follow the tutorial for more details on how to use the tool and how to interpret its output.

Running the Experiments

The experiments provided in the examples folder represent those that are still functional and were used as proof of concepts into previous papers this project was involved. For a step-by-step tutorial on how to setup your own experiment, check out the tutorial provided in this repo.

Included examples

• DSD18: experiments from our DSD'18 dealing with TDN NoCs exploration that optimize power while respecting real time constraints.
• ScalAnalysis: folder containing scripts that generates experiments for different sized NoCs platforms based on a template extracted from DSD18.
• tutorial: the files used for the user tutorial.

Publications

Kathrin Rosvall, Tage Mohammadat, George Ungureanu, Johnny Öberg, and Ingo Sander. “Exploring Power and Throughput for Dataflow Applications on Predictable NoC Multiprocessors,” 719–26, 2018.

Kathrin Rosvall, Nima Khalilzad, George Ungureanu, and Ingo Sander. Throughput propagation in constraint-based design space exploration for mixed-criticality systems. In Proceedings of the 2017 Workshop on Rapid Simulation and Performance Evaluation: Methods and Tools (RAPIDO '17), Stockholm, Sweden. ACM, January 2017.

Nima Khalilzad, Kathrin Rosvall, and Ingo Sander. A modular design space exploration framework for multiprocessor real-time systems. In Forum on specification & Design Languages (FDL '16), Bremen, Germany. IEEE, September 2016.

Kathrin Rosvall and Ingo Sander. A constraint-based design space exploration framework for real-time applications on MPSoCs. In Design Automation and Test in Europe (DATE '14), Dresden, Germany, Mar. 2014.