Curie Cluster
Curie is specifically configured for running computing jobs with parallel Gaussian 03.
User Accounts
To get an account to run jobs on Curie, please email Hieu Tran (tran at chem wisc edu).
Usage Policies
Failure to observe prescribed usage policies may result in the unceremonious termination of any and all jobs in violation.
- Each user is allowed to use a maximum of 8 processors (4 nodes) total for all concurrent jobs at any given time. In other words, if you run a large job using 6 processors, you can run an additional job using 2 processors, but no more.
- Do not submit jobs on Curie directly. Always use the queue to submit jobs
- If your job has crashed and you cannot clear it, please email Hieu Tran (tran at chem.wisc.edu) immediately so we can assist you and/or restore the queue.
Access
Users must navigate to nbohr and curie via Secure Shell Protocol (SSH). Check out the links for several software that supports SSH, available on the remote access page.
Access to Curie involves two steps (example shown for user 'tran'):
Step 1: Connecting from your local computer to nbohr:
- Linux and Mac OS X users: Open a terminal, use the following command to connect to nbohr (example shown for user 'tran'):
[mycomputer]$ ssh tran@nbohr.chem.wisc.edu
- Windows users can use remote access software such as Putty or SecureCRT to access nbohr. For more info, go to the remote access page.
Step 2: Connecting from nbohr to curie:
After logging on to nbohr, users can connect to curie using the following command (example shown for user 'tran'):
[nbohr]$ ssh tran@curie
Transfering files from/to Curie:
Users must navigate to nbohr and curie via Secure Shell Protocol (SSH). Check out the links for several software that supports SSH, available on the remote access page.
Transfering files to Curie involves two steps (example shown for user 'tran'):
Step 1: Transfering files from your local computer to nbohr:
- Linux and Mac OS X users: Open a terminal, use the following command to transfer files to nbohr (example shown for user 'tran'):
[mycomputer]$ scp localfile tran@nbohr.chem.wisc.edu:/home/tran/remotefile
(change 'localfile' and 'remotefile' to suit your needs)
- Windows users can use file transfering software such as SecureFX to transfer files to nbohr. For more info, go to the file transfering page.
Step 2: Transfering files from nbohr to curie:
After logging on to nbohr, users can transfer files to curie using the following command (example shown for user 'tran'):
[nbohr]$ scp localfile tran@curie:/home/tran/remotefile
(change 'localfile' and 'remotefile' to suit your needs)
Running and Monitoring Jobs on Curie
Running Jobs:
Always use the queue to run jobs on Curie. Jobs submitted outside the queue will be terminated. An example is shown below.
[curie]$ qg03 [NP] h2o.g03
(i.e. read Q-G-Zero-Three [NP] ...)
Where: [NP] = 1,2,4,6,8 - number of processors to be used
Monitoring jobs (Checking the queue status):
To check the queue status, type:
[curie]$ qstat -n
Available Software
Curie is specifically configured for running computing jobs with parallel Gaussian 03.
Hardware Configuration
File Server
- Processors: dual, Intel Xeon, 2.8 GHz
- Memory: 1GB
- Hard drives: 250GB
Individual Nodes
There are a total of 29 nodes. Specifications are shown below.
- Processors: dual, Intel Pentium III, 800 MHz, 256 KB cache
- Memory: 1GB
- Hard drive: 160 GB
Citation
If you have used the Curie cluster for your research, please visit our citations page.