Getting starting with the Roomful of Mac Minis

There are 28 Mac Minis located in the APM 2402 Classroom.
These machines are accessible remotely.

These machines have a low-end Nvidia GPU that will enable
to you some of your code development, in particular,
concerning correctness, but only some basic  performance tuning.

The 28 machines have the following names (note that there
is NOT a leading zero for machines 1 through 9)


 csme-apm2402-1.ucsd.edu to csme-apm2402-28.ucsd.edu.

To increase the likelihood of quickly finding an available machine,
use the following algorithm.  Take the day of the month that you were born,
and compute its value modulo 10.  Pick a machine based on that number.
So if you were born on the 29th day of the month, choose
 csme-apm2402-9.ucsd.edu


When you login, check to see if anyone else is logged
in using the "who" command.  If so, then try another machine,
by adding 10 to the machine number you first chose, i.e. for the 
above example
 csme-apm2402-19.ucsd.edu

Repeat again if necessary or pick a machine at random.
If the sum exceeds 28, e.g. 29, then choose  29-28=1:
  csme-apm2402-1.ucsd.edu

Note that, between the hours of 6PM and 2AM, you may use these machines at will.
At 3AM each the machine may be unavailable for several hours of maintenance
Classes are held in APM 2402 during the day, and students may use the machines.
Keep this in mind.  Here is the schedule for the room

http://www.math.ucsd.edu/announcements/seminar-room-schedules/apm-2402/

If you see someone logged in when a class is in session, who is not
in our course, log off immediately and try another time when the class
has ended.

Due to security and other concerns, the Mac Minis have  been
assigned IP addresses that are routable within UCSD only, for both
incoming/outgoing network traffic.  Thus, to access the Mac Minis,
you must either be logged into a machine on campus having
full internet access, OR use VPN. However, outgoing connections
are not possible. So, if you wish to transfer files to Dirac, you must
do so through an intermediary campus machine.

When you login the first time, go to  the directory
/share/class/public/cse260-fa12/Profiles

Append a copy of the file BASH_PROFILE to the end of your .bash_profile

cat BASH_PROFILE >> ~/.bash_profile

Next, logout and login again. Your environment will now include the
familiar $PUB variable which points at the above path

You can find the increment array example discussed  in class in
the following directory

$PUB/Examples/CUDA/incrArr

Copy that directory your $HOME, cd to incrArr and type "make"

HARDWARE CONFIGURATION

All the Mac Minis are configured as follows:

  HARDWARE:
  Apple Mac mini
  Intel Core 2 Duo 2.66GHz CPUs
  8GB 1067MHz DDR3 RAM
  320GB HD
  NVIDIA GeForce 320M 256MB integrated video graphics
  (provides 48 CUDA cores running at 0.95GHz)

  SOFTWARE:
  Apple Mac OS X 10.7.5 Lion
  NVIDIA CUDA Driver 5.0.24 for Mac
  NVIDIA CUDA Toolkit 4.2.9 for Mac
  NVIDIA GPU Computing SDK 4.2.9 for Mac
  Apple Xcode 4.2.1 compiler installed with the Command Line Tools of
  "gcc version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.1.00)"