Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
72 lines (61 sloc) 3.13 KB
CMPT 431 Antix Project instructions
0. rtv (instructor) todos:
- collisions
- scoring
1. Prerequisites
- a UNIX-like OS (tested on Linux and OS X)
- OpenGL and GLUT
- C++, STL, Make
2. Obtaining the code:
- The original Antix repository is at
- learn about github at
- learn about git at
- Cloning the repo (to get your local copy to work on)
- git clone
3. Building and running the code
- cd Antix
- (edit the Makefile to suit your system: OS X and Linux setups are provided)
- make
- ./antix
4. The Big Task
- Extend (or reimplement) Antix to support
- 1,000,000 robots
- the simulation backend MAY be implemented as multiple threads, processes or hosts
- 1000 homes (teams) each with at least one independent controller
- controllers MUST be separate OS processes
- controllers MAY run on separate hosts to the simulation backend
- a visualization of the whole world
- the system should run as fast as you can make it, given the time available.
- your target platform is the CSIL laboratory. Additionally, access to the CMPT cluster or
Amazon Elastic Compute Cloud can be arranged privately or through the instructor.
- using git for SCM is _strongly_ recommended. You may find github useful for collaboration, too.
5. The First Task
- Design a system architecture and development plan that can achieve the Big Task
- no code, just design
- split your team equally between backend and client-side responsibilities
- present a technical argument for your design
- why will it work?
- what could go wrong?
- what is Plan B, C?
- what remains to be decided?
- present a plan for development and testing, considering balanced workload for team members
- presentations will be by 1 person from each team
- 20 minutes
- use computer, transparencies or blackboard BUT
- your slides and notes must be submitted on the day, and should be comprehensible by themselves
- Presentations will be in class on tuesday 25 January and followed by class critique.
6. Subsequent tasks
- depend on your development plan, but each team will report each week, alternating between
backend and client sub-teams. Every person presents at least once.
- Hint: improve single-thread performance by improving data structures, etc. before using
parallelism. This exercise will also help you understand the problem and the code.
- The final week of semester and the exam period will be an AI contest using one or more
of the online Antix systems you build.
8. Getting help
- Use the github systems to submit bug reports, pull requests, etc.
- Your TA is Abbas Sadat [], and Autonomy Lab PhD student with relevant experience.
You can ask him anything except to do your project for you.
- Use the class mailing list for questions and discussions.
- Ask Abbas and Richard questions via the mailing list so we can avoid duplicates, etc. Please
reserve personal email only for private business.