In ITS News No.209, we announced the new addition of the AI-Research computing platform to the University's research computing resources. Today, we are glad to announce that the following systems for supporting research have formally been put into service:
The HPC2021 (hpc2021.hku.hk) cluster is a High Performance Computing (HPC) system equipped with hundreds of compute nodes giving an aggregate theoretical processing power of 804 TFLOPS. Besides traditional compute-intensive jobs, there are computing nodes each with 128 CPU cores and up to 2TB memory to support memory-hungry applications such as bioinformatics analysis, and computing nodes with up to 8 V100 GPUs for running AI/deep learning jobs at scale (e.g. training and inference on large-scale datasets). Backed by a 1.2PB Lustre parallel file system, it is a good system for research simulations which incur large input/output files.
For details, please refer to the HPC2021 website at https://hpc.hku.hk/hpc/hpc2021/.
The AI-Service system is a graphical user interface (GUI) based integrate Artificial Intelligence (AI) system built with the IBM Watson Machine Learning Accelerator (WMLA). Users working on research studies in different disciplines can have access to the AI software/tools readily available on the WMLA platform for doing data preparation, model training and inference deployment for their AI/Machine Learning (ML) application developments, mostly via a web interface.
Users who would like to know more about the AI-Service system can join the following two online training sessions organized by IBM:
Date: May 14 and May 17, 2021 (2 half-day sessions covering different topics)
Time: 9:30 am - 12:30 pm
May 14, 2021 session – https://hkuems1.hku.hk/hkuems/ec_hdetail.aspx?UEID=75068
May 17, 2021 session – https://hkuems1.hku.hk/hkuems/ec_hdetail.aspx?UEID=75069
More details will be updated in the AI-Service website at https://hpc.hku.hk/ai/ai-service/.
The AI-Research system is based on a nVidia DGX A100 server to facilitate sophisticated AI research with the use of most advanced A100 GPUs. The AI-Research system was launched earlier in January 2021. Configurations of the server system allow better user-system interactivity in editing/developing a new AI/ML pipeline or neural network architecture. The users are required to handle software installation on their own with the use of software container technology. Researchers can visit https://hpc.hku.hk/ai/ai-research/ for more information.
Research Computing Team
Phone: 3917 2485
Email : firstname.lastname@example.org