The AI and Simulation Lab (ASL) Server provides a powerful platform for conducting high-performance computing tasks, enabling researchers and students to run computationally intensive jobs on advanced GPU and CPU resources.
Server Access
To access the ASL server, please use the SSH command below:
ssh -p 44065 [Your Username]@203.159.10.44
Ensure that your user credentials are correct and that your network settings allow SSH connections to the server.
Remark: In case you want to access ASL Slurm from off-campus, please make connection to AIT-VPN before access ASL.
System Specifications
- ASL-gpu (Default Partition): Equipped with 4 x NVIDIA A6000 GPUs, ideal for deep learning and GPU-accelerated computations.
- ASL-cpu: Features 128 V-CPUs, suitable for CPU-based processing and parallel computing tasks.
Getting Started with Slurm
Slurm is the job scheduling system used on the ASL server to manage and allocate resources efficiently. To submit and manage jobs, follow these essential steps:
- Creating a Slurm Batch File
- Follow this guide on creating a Slurm script to write and submit batch files effectively.
- Tutorial Video
- Watch the comprehensive tutorial on how to use Slurm in the AI and Simulation Lab: YouTube Link by Dr. Chaklam.
Additional Tips
Familiarize yourself with basic Slurm commands such as
sbatch,squeue, andscancelto manage your jobs.Make sure your batch script specifies the appropriate partition (e.g.,
#SBATCH --partition=ASL-gpu).
With these resources, you will be able to maximize your experience using the ASL server for your research and projects. For further assistance, contact your lab administrator or refer to the user manual provided by the AI and Simulation Lab.