|May’20-Round||4 May 2020 to 30 June 2020||Closed|
On 12 March 2019, Mr Heng Swee Keat, Minister of Finance and Chairman of the National Research Foundation (NRF) Board, announced that S$200 million will be allocated by the NRF to support the upgrading of Singapore’s supercomputing infrastructure. As a national research infrastructure, NSCC will be extending these supercomputing resources to local research institutes, companies and institutes of higher learning (IHLs), which include autonomous universities and polytechnics. While education-based curriculum and projects have access to wet-lab experiments, high performance computing (HPC) provides immense new capability to perform larger dry-lab experiments beyond available on-premise or chargeable cloud resources. In addition, HPC support introduces a new dimension in experiential learning by providing students with new opportunities to explore larger and more complex problems that are beyond typical textbook examples. The call aims to strengthen module delivery and education by leveraging unique and sought-after HPC resources. Teaching modules of interests include but are not limited to science and engineering (eg. in areas such as modelling & simulation), AI, machine learning and big data analytics. Possible uses of educational HPC resources include supporting lab exercises & projects in teaching modules, project-based modules and competitions like Datathons, among others.
Currently, all staff and students from NSCC stakeholder organisations (A*STAR, NTU, NUS, SUTD, TCOMS, NEA) can self-register for an NSCC account with 100,000 core-hours and 50GB of storage as default quota. This Educational HPC Projects call, endorsed by NSCC Steering Committee chaired by Mr Peter Ho, is an opportunity for all IHLs to request for additional HPC resource quotas of up to 1M core-hours and 0.1 TB of storage. To support graduate studies, graduate students can apply for resources to run large compute-intensive experiments. To help you get started on HPC resources, NSCC has developed a pool of user guides, supporting materials and training workshops on how to access and use NSCC’s resources.
3. Type of Resources Available
- ASPIRE 1, a petascale supercomputer, which consists of 1,288 nodes of CPU and accelerator nodes with NVIDIA K40 GPUs
- AI System, which consists of 6 units of state-of-the-art 8-GPU NVIDIA DGX-1 with V100 cards. This system is meant for projects that focus on AI research with particular novelty on scale and/or throughput breakthrough
- 13PB of High Performance Storage
- Software: This Call for Projects will also take into account applicants’ software requirements. NSCC will consider installing open-source software if not already available in the system or facilitate the bulk procurement of popular commercially available HPC teaching and training software applications.
Applications will be reviewed by an Educational Projects Selection Committee comprising HPC experts from both academia and industry. Approved resources will have to be utilised within one year from the time of allocation, wherever possible. As this is the first call, and in view of potential disruption to regular IHL schedules and processes due to COVID-19, NSCC is open to extensions if required.