Hank HOWTO - Running the Simulator

This document pertains to the execution of Hank.
Last Update:

Quick Info

Run Hank. At the moment, you must run hank on either corn or tractor. First change directory to the hank/src/main (assuming the hank directory is located in your root directory.)

The form of the execution line is:

hank <scheduler file> <edf file>

The <scheduler> file specifies the processes to be run and the order in which they are to be run. The <edf file> specifies the road database which defines the layout and logic of the virtual environment. As an initial test, try the following example:

hank ../../data/sp_hank ../../data/edf_files/biketown.edf

This invokes a sequential schedule that does not include a visual process. You can run it while remotely logged in to corn or tractor. The only output will be text printed to your screen.

For a second test, try running the following command from the corn console. It will display a 2-D rendering of the simulation.

hank ../../data/sched_sdl_vis_seq ../../data/edf_files/biketown.edf

Runaway Processes

Occasionally, the hank code will leave behind hank threads due to an unexpected crash of the system. Periodically, you may want to check the process listings and try to determine if you have any rogue processes on the machine.

The leftover hank threads will show up in the process listing as hank. There may be several of them. If you want to scan the processes on an SGI, here's a simple command that can be used to view the process listing according to largest size process.

ps -elf | grep -v "SZ:RSS" | sort +9 -r

Similarly, if you only want to see the top 10 items,

ps -elf | grep -v "SZ:RSS" | sort +9 -r | head -n 10

Or, if you only want your processes, try

ps -fu <username>

To remove all of the hank threads, use the following command:

/sbin/killall hank