HPC FAQs
What is High-Performance Computing (HPC)?
HPC involves the use of supercomputers and parallel processing techniques to solve complex computational problems that require significant processing power and memory.
How do I get access to an HPC system?
Currently HPC resources are available to faculty members and research students of the University. Application can be made through the online ITSC/Library helpdesk system.
What types of applications are best suited for HPC?
Applications that require large-scale simulations, complex calculations, data analysis, and modeling, such as weather forecasting, genomics, and computational fluid dynamics.
How do I submit a job to the HPC system?
Jobs are usually submitted through a job scheduler like SLURM, using scripts that specify resource requirements and execution commands.
What is a job scheduler, and how does it work?
A job scheduler manages the allocation of HPC resources, queuing user jobs, and optimizing the execution of tasks based on resource availability and priority.
What is parallel computing, and why is it important in HPC?
Parallel computing involves dividing a problem into smaller tasks that can be executed simultaneously across multiple processors, significantly speeding up computation.
How do I optimize my code for HPC systems?
Optimization can involve parallelizing code, using efficient algorithms, minimizing I/O operations, and leveraging hardware-specific features like vectorization.
How do I manage data storage and transfer on an HPC system?
Data management involves using high-speed storage systems, efficient file I/O practices, and data transfer tools like SCP for moving data between your desktops and the HPC platform.
How do I troubleshoot job failures on an HPC system?
Troubleshooting involves checking job logs, verifying resource requests, debugging code, and consulting system documentation or support staff.
What are the best practices for using HPC resources efficiently?
Best practices include optimizing code, requesting appropriate resources, minimizing job run times, and using job arrays for parameter sweeps.
How do I ensure the security of my data on an HPC system?
Security measures include using strong passwords, enabling encryption for data transfer, and following institutional policies for data access and storage.
Where can I find HPC documentation and user support?
Documentation and user support are available through the ITSC website and online ITSC/Library helpdesk system.