The STATA Data Analysis package (http://www.stata.com/) is not installed centrally. However, the University of Iowa provides a pathway to an academic license so users may purchase a copy for individual use. See this page for details: http://cs.its.uiowa.edu/software/stata.shtml.
Once you have acquired a copy of STATA, and want to run it from your home directory, you will need to use the following steps.
- Get your copy of STAT up to the cluster system. You'll need to copy the entire contents of the STATA DVD into a folder in your home account. You can achieve this by either mounting your home drive on a local computer that has a DVD drive (see: Home Accounts for instructions), or use sftp or Globus Online (see: Globus Online) to transfer the files.
- Once you have the STATA install software in a folder inside your home account, you will need to create another folder that will contain the installed product. Log into your account (if you are using a desktop login, open a terminal window). For example purposes, assume install files are in ~/stata-install.
Code Block | ||
---|---|---|
| ||
[user@helium-login]:$ mkdir stata [user@helium-login]:$ cd stata [user@helium-login]:$ sh ../stata-install/install |
3. Next, you will see installer prompts. Follow the prompts to install Stata into the ~/stata directory. Make sure the install directory (listed as "Current Directory") is the directory you created above. Below are example screenshots of an install:
Next you should see the installer working. Once the install is finished, you will need to activate the software by executing:
Code Block | ||
---|---|---|
| ||
[user@hpc-login]:$ ./stinit |
Enter the license information you received when you purchased STATA. Finally, you can add the freshly installed STATA program into your path variable so that it's executable from outside your ~/stata directory. To do this, edit your ~/.bash_profile directory, and edit the path line as below:
Code Block | ||
---|---|---|
| ||
PATH=$PATH:$HOME/bin:$HOME/stata |
Info |
---|
You can also create your own environment module to manage this. |
When you log in again, you'll be able to simply execute stata or statax:
Code Block | ||
---|---|---|
| ||
[user@hpc-login]:$ stata |
While it is possible to run STATA interactively, which can be especially useful for prototyping a job, please remember that login nodes are a shared resource and are intended for launching jobs or prototyping smaller versions of jobs you intend to run. It is not advisable to use interactive sessions for long-running, compute-intensive jobs unless one uses a qlogin session to do so.
You can run STATA in batch mode, and there is information on doing so here: http://www.stata.com/support/faqs/unix/batch-mode/. Note that this assumes you are familiar with using STATA's do-files, which are scripts with series of STATA commands that you can feed to STATA to run automatically. To become more familiar with what do-files are and how to use them, please consult: http://www.stata.com/manuals13/u16.pdf. You'll first want to build a STATA do-file, then submit that to SGE using qsub. Please refer to the following pages on submitting jobs for how this works, and also for how to request needed resources: Basic Job Submission and Advanced Job Submission. Note that not all versions of STATA are MPI-enabled, so depending on which version you have purchased, you may or may not be able to make use of parallel processing in STATA.