PARAM Himalaya Supercomputing Facility

PARAM Himalaya Supercomputing Facility at IIT Mandi provides high-performance computing resources for scientific research, engineering simulations, artificial intelligence, machine learning, computational modelling, and data-intensive applications. The facility is managed by Digital & Computing Services (DCS) and enables researchers to execute parallel applications efficiently using the SLURM workload manager.

Researchers can access CPU nodes, High Memory nodes, and GPU accelerated nodes for their computational workloads.

Visit PARAM Himalaya Website

System Information

156

Total Nodes

107

CPU Nodes

39

High Memory Nodes

10

GPU Nodes

PARAM Login Information

Connect to PARAM Himalaya using SSH: 

ssh yourusername@paramhimalaya.iitmandi.ac.in -p 4422

Login Procedure
  1. Open Terminal and connect using the SSH command shown above.
  2. Complete the CAPTCHA verification.
  3. Scan the QR code using the Google Authenticator app.
  4. Register your account in Google Authenticator.
  5. Enter the 6-digit OTP generated by the app.
  6. Enter your PARAM password.
  7. You are now successfully logged into the PARAM Himalaya Supercomputing Facility.

Transferring Files using rsync

If you need to transfer files between your local computer and the HPC cluster, it is recommended to use **`rsync`** instead of **`scp`**. `rsync` is faster, more reliable, and transfers only the portions of files that have changed, making it highly efficient for large datasets and repeated synchronizations. It also supports resuming interrupted transfers, allowing data transfer to continue from where it left off in the event of a network failure or other interruption.

Copy files from your local computer to the HPC:

rsync -avz /path/to/local/file_or_directory
<username>@paramhimalaya.iitmandi.ac.in:~/destination/

Copy files from the HPC to your local computer:

rsync -avz <username>@paramhimalaya.iitmandi.ac.in:~/source/
/path/to/local/destination/
Note: Replace <username> with your HPC username and update the source and destination paths as required.

SLURM Related Information

All compute nodes are connected using a High-Speed InfiniBand (IB) network.

Command Description
srun Run a command on allocated compute nodes.
sbatch Submit a batch job.
squeue Show status of queued jobs.
sinfo Display compute node status.
salloc Allocate nodes for interactive jobs.
scancel Cancel submitted jobs.
#SBATCH -A account_name Specify project account in your job script.

Module Files

Users are strongly encouraged to use the pre-installed software modules available on the PARAM Himalaya cluster.

module avail

module load <module_name>

module unload <module_name>

module list
Guest Users

Guest users are invited to experience the advanced capabilities of the PARAM Himalaya Supercomputing Facility. Access for external users is available on a paid basis as per the institute policy.

PARAM Guide 1

Facility

PARAM Guide 2

Charges

For any queries, kindly email us at it_helpdesk@iitmandi.ac.in