Running SAS®, SPSS, and BMDP on Zoo
Running SAS® and SPSS on the Academic
Zoo Cluster
The SAS System and SPSS are two of the most widely used programs
for statistical analysis at UVM. Both are available to any UVM affiliate
with an account on the academic Zoo cluster. Both can be run in
any one of four different modes of execution, depending on the user's hardware
and/or on the user's personal preference.
X Windows Interactive Mode
X Windows provides a graphical user interface to many applications running
on Unix systems. SAS and SPSS can be run on an X terminal or from
a microcomputer running an X server application. This is the mode of execution
expected by default when you type "sas" or "spss" at the system prompt.
If your hardware is X Windows capable, you will be presented with a set
of interactive windows, allowing you to navigate through the programs using
a mouse. This mode of execution provides the highest level of interactivity
with the programs, similar to that provided by running them on a Macintosh
or in Windows. Refer to the respective product manuals for information
on how to proceed from here.
ASCII Terminal Line Mode
This mode of execution allows for interactive use, but instead of a window
in which to interact with the program, a line prompt allows you to input
commands one after another. Program output is displayed on your screen.
The commands to invoke SAS and SPSS in this mode are as follows:
sas -nodms
spss -m
No output files are created when the programs are used in this manner.
Non-interactive Mode
Both statistical packages can be run in a non-interactive mode. The
programs accept commands from a file, execute those commands, and place
the results into an output file. You should have an existing file with
a .sas or .sps extension containing your commands. The file can be
created by using a text editor such as pico or emacs or may be created
by a word processor and transferred from your microcomputer. Data may be
included in the same command file, or a separate data file may exist which
is then referenced by the SAS or SPSS command file. Execution begins
after issuing the following commands and ends when the system prompt returns.
sas filename.sas
spss -m filename.sps > filename.lst
SAS will read the commands in filename.sas and create two new files; one
called filename.log containing SAS informational notes and error messages,
and one called filename.lst containing the actual SAS procedure output.
If filename.log and/or filename.lst already exist, SAS will automatically
overwrite these files. SPSS will read the commands in filename.sps and
redirect the output to filename.lst. If filename.lst exists and you have
set the environment variable
noclobber, the output file will not
be overwritten. In order to overwrite the output file, you must issue the
command
unset noclobber at the system prompt before running SPSS.
Batch Mode
Batch execution is the preferred method if your SAS or SPSS job
is likely to consume a great deal of CPU time. In order to run jobs in
batch, you will need to have already created SAS or SPSS command
files as described in the previous section. You should also create a file
containing the commands you want the batch processor to execute. For example,
if you want to run a SAS program in batch, put the command
SAS filename.sas
into a file called, say, run.batch. Make this file executable by typing
chmod a+x run.batch.
Then submit this file to the batch processor by typing:
submit run.batch
at the system prompt. To find out whether you have any batch jobs running
use the command
jobs at the system prompt.
Last modified April 10 2008 02:56 PM