Folding@Home Monitor Application for
Linux
download the latest version here
screenshots here
CONTENTS
I. What Is Protein Think?
II. Quick Start
III. Installation and Setup
IV. Usage
V. Known Problems
I.
WHAT IS PROTEIN THINK?
Protein Think is a program that
monitors the progress of a the Folding@Home client, running on a *nix
platform. It displays data and statistics in a straightforward manner,
as well as allowing users to put personal or team graphics on their
user interface. While PT attempts to present information in a
straightforward manner, if you do not know what the information means,
it might not be of much interest. Much of the rest of this document
presents a very brief explanation of information that PT presents on
screen.
II. QUICK
START
Assuming you have a running
Folding@Home client, just:
1) Run the setup script ('python
fahsetup.py'), and
2) Run the application ('python
fahmain.py')
3) There is no third thing
III.
INSTALLATION AND SETUP
Before running Protein Think,
you should have a working Folding@Home client on your machine. PT
simply monitors the output of a F@H client, so without one PT has
nothing to do. You can download a F@H client at:
http://folding.stanford.edu/download.html
If you are running ubuntu Linux,
there is a script that will download and setup your client here:
http://www.ubuntuforums.org/showthread.php?t=101817
Assuming you have a F@H client
running, the next thing is to wait for it to produce its first
unitinfo.txt file. This can take a few hours, depending on the speed of
your machine and the difficulty of the protein simulation you have been
given. Protein Think won't allow you to complete the setup until the
unitinfo.txt file exists.
Once the unitinfo.txt file has
been born, you are ready to setup Protein Think. Untar the tar.gz file
to any handy location (for example:
/home/<username>/protein_think). Then open a terminal shell and
navigate to that directory (for example: '$ cd
/home/<username>/protein_think').
Now run the setup script, by
entering: 'python fahsetup.py'
The setup script will lead you
through the rest of the setup process. In order to complete this
process, you will need to know the absolute path to the unitinfo.txt
file. You will also need to know your F@H username (created when you
set up the F@H client) and team number (if you are not on any team,
your number is 0).
The setup script will create a
hidden directory called '.stats' will PT will use. The simplest place
to put this directory is in the protein_think directory (for example:
/home/<username>/protein_think/.stats), however you can put it
anywhere you like.
Once you have finished with the
setup, you are ready to run Protein Think.
Protein Think gets its
information about the activities of your Folding@Home client from two
sources: your computer and the Folding@Home website. Therefore it helps
if you are connected to the internet. If you are not connected, PT will
still run, but several of its data pages will be blank.
To start the application, open a
terminal and navigate to the protein_think directory (for example:
/home/<username>/protein_think).
Then enter 'python fahmain.py'
The program will then launch.
IV.
USAGE
Protein Think displays its
information in a notebook-style layout with several pages. A
description of each page follows.
- The Work Unit Page
The Work Unit Page ("WU") is
probably the simplest page in PT. It shows the name of the protein
being currently being folded, and how much of this work unit has been
completed (expressed as a percentage). The Download Time and Due Time
fields are self explanatory, except for one note: These are expressed
in UTC, or roughly Greenwich Mean Time, which is probably different
than the local time on your computer, unless you live in, say,
Greenwich, England and environs. ;)
- The Individual Statistics Page
The Individual Statistics Page
("Indv") includes some numbers that are potentially confusing,
especially if you are not on a team. The name, credits, and work units
fields are self-explanatory, unless you are totally unfamiliar with
Folding@Home. In a nutshell, credits and work units are two different
ways that Folding@Home keeps track of how much folding your client has
done for them. Work units are not all the same size -- some proteins
are much more complex than others -- so some work units count for more
credit than others. For example, a work unit that takes a computer two
days to complete may be worth fewer credits than a work unit that takes
the same computer a week to complete.
The "Indv Rank" field is your
standing among all folders, regardless of whether they are on a team or
not. The "Rank on Team" field is your standing on your folding team.
The wrinkle here is if you are not on a team. Individual folders are
automatically assigned to the default team, which is team number 0. So,
even if you are not on a team, you have a de facto standing on team 0.
In other words, this is your rank among all the lone folders.
If you want to see more detailed
statistics, go to this page: http://folding.stanford.edu/stats.html
- The Team Statistics Page
The Team Statistics Page
("Team") is straightforward unless, again, you are not on a team. It
displays the name of your team, the team number (which is the team's
actual name), the work units and credits your team has amassed, and
finally the standing of your team among all teams. If you are a lone
folder, then your team is "Default" and the work units and credits are
the sum total of all those churned out by individuals on team 0.
Interestingly, team 0's rank is always 0.
The data on both the Individual
and Team Statistics pages are downloaded from the appropriate file on
the internet. This file is updated only a few times a day, and PT only
checks every few hours, so when your client finishes a work unit do not
expect your work unit and credit scores to jump immediately. After
finishing a work unit, it usually takes 4-6 hours before PT will
display the updated work unit and credit scores.
- The Queue Data Page
The Queue Data Page ("Queue")
presents some very detailed information that is normally stored in a
binary file in the F@H client work folder. The late Dick Howell created
a program in C (called "qd.c") that reads this file and outputs its
contents in human-readable form. I have provided a copy of that program
in the PT distribution, and formatted the output very slightly to
present it in a text window. Delving into an explanation of the meaning
of the arcana presented on this page is beyond my self-motivational
skills right now. However, if you want to look further into what it's
all about, this is a good place to begin:
http://home.comcast.net/~wxdude2/rph/fah.html You can also download qd
tools from: http://linuxminded.xs4all.nl/?target=software-qd-tools.plc
III. KNOWN
PROBLEMS
PT is dependant on the
unitinfo.txt file to report the progress of the current work unit. A
very few work units don't use a unitinfo.txt file, or use it in an
alternate manner, and this might break PT for the duration of that work
unit's run. Part of the reason PT isn't set up to handle this, is that
I have yet to see one of these rogue work units on my machine.
Protein Think is hacked together on ubuntu.
copyright © Michael
W. Taft, 2006