November 15, 2005 - If you are having problems running Medici on Assignment 4, it is likely a problem with the configuration of X-Win. As a workaround, search through your *.inp files for PLOT commands, i.e., PLOT.1D and PLOT.2D, and add a line something like:
+ DEVICE=CP/POSTSCRIPT PLOT.OUT=<filename>.ps
immediately after each PLOT line, where <filename> is the name of a Postscript file where you would like your image saved. You may then view the images using gv as on previous assignments.

November 14, 2005 - On Assignment 4, when computing the flatband voltage, please ensure that your units work out correctly, i.e., you may need to multiply QSS by the magnitude of the electronic charge in order to get units of volts.

October 27, 2005 - An image of the world's highest f_T HBT posted in the downloads section under IMAGES.

October 4, 2005 - If Medici is giving you a "Killed" error, do the following:
(1) When you login via SSH, type: echo $DISPLAY
(2) You should see something like ug1.ece.ubc.ca:XX.0, where XX is a number.
(3) If XX is too high, there are too many users with X11 tunneling connections on the present machine; thus, you need to login to a different undergrad (ug) computer.
(4) To switch machines, simply type at the current prompt "ssh ugYY", where YY is a number from 1 to 10 that is different from the ug machine you started out on. Once logged in to the ugYY machine, simply use Medici as usual.

October 3, 2005 - Note that Medici is accessible from the Solaris and Linux machines. The Solaris computers may be accessed remotely by SSHing to ssh.ece.ubc.ca. From there, follow the directions in the Medici tutorial detailing how to configure your account appropriately. If you wish to use a Windows machine, you will need SSH and, most likely, X-Win32. See the instructions in the downloads section if you need to configure these programs. If you wish to use a Linux machine, you will need to modify your .cshrc.linux file instead of the .cshrc.solaris one. The appropriate modification on Linux is to add the line "set path = ($path /CMC/tools/tcadlnx/bin)" (without the quotes). If you are using Solaris, either directly or through SSH, you must add "set path = ($path /CMC/tools/tcad/bin)" (without the quotes) to your .cshrc.solaris file as specified in the Medici tutorial. The changed file will not take effect until you issue the "source <filename>" command, or you logout and login again. Files may be edited with the text editor of your choice. Also, note that your *.inp files must be accessible by the machine that is actually running Medici. If using Windows, it may be most convenient to place the files on your Z: drive since the Solaris machines may access these folders directly. Some basic Unix commands may be found here. If you are having troubles, please contact the TA.

October 3, 2005 - Note that *.pdf documents may be opened, on the Solaris and Linux machines, by executing the command "acroread <filename> &" without the quotes, where "<filename>" is the file you wish to view. The "&" at the end of the command allows you to continue to access the command line while Acrobat Reader is active. Similarly, *.ps and *.eps files may be viewed by executing the command "gv <filename> &".

September 7, 2005 - email from Mr. Tootoonian re. brain power of lizards:
In class today you showed that slide by Hans Moravec plotting the progression of computing power in terms of various biological organisms. I'm quite into AI and computational neuroscience (I'm planning on pursing the latter in grad school), so I have a pretty good idea of how he came up with the figures that he did. It's a bit hokey, but I thought you might be interested anyway. You mentioned that you've done some artificial retina research, so I do apologize if you already know all of what follows. There are roughly 1011 neurons in the human brain, each making on average 104 synapses (though some, like the Purkinje cells in the cerebellum make around 200000!!). The neurotransmitters released in these synapses take on the order of 0.1 seconds to cross the synaptic cleft and have their desired effect on the receiving dendrite. Multiply all these numbers, and you get 1016 'operations' per second. These operations are then measured against MIPS numbers for modern computers and you end up with graphs like Moravec's. The only problem is that there are so many qualitative differences between how brains work and how computers work that such comparisons don't really make any sense. A good analogy I've come across that captures this is that if you sped up the brain of a dog a million times, it wouldn't suddenly start talking and writing equations.