Co-located with Supercomputing Asia
When: 24th Feb 2020
Where: Suntec Singapore Convention and Exhibition Center
Altair PBS Works™ is the market leader in comprehensive, secure workload management for high-performance computing and cloud environments.
A rapidly growing number of organisations in Asia have adopted PBS Works technology, and Asia now joins the US as a PBS Works User Group meeting hub. Join us in Singapore on 24 February to hear about PBS Works latest features and solutions, catch up on Altair activities and partnerships, and network with your peers.
PBS Works User Group meetings are a two-way channel. We learn as much from you as you do from us, and we’ve made product improvements based on your feedback. Now with a critical mass of PBS Works users in Asia, you have the opportunity to tell us what features will help your business.
This event is by invitation only. We’ll cap it off by recognising some key PBS Works contributors and users who’ve accelerated the adoption of HPC and PBS Works in the region.
Schedule
| 0900 – 1000 | PBS Works User Group Inauguration and road-map Bill Nitzberg - CTO, PBS Works |
| 1000 – 1020 | HPC reliability improvement and container integration with PBS Works SOH Hwee Jin, Melvin - Senior Assistant Director - High Performance Computing Centre, NTU Edwin Tan Seng Tat - Senior Assistant Manager - High Performance Computing Centre, NTU |
| 1030 – 1100 | Morning Tea Break |
| 1100 – 1120 |
GPU monitoring and management with NVIDIA data center GPU manager (DCGM) Dr. Gabriel Noaje - Senior Solutions Architect, NVIDIA APAC |
| 1120 – 1150 |
PBS Works for HPC and AI Subhasis Bhattacharya - Senior Director, Software Development |
| 1150 – 1220 | User Group Open Forum Discussion An open-forum discussion to discuss and share challenges, ideas and best practices for PBS Works. |
| 1220 – 1230 |
Closing Notes and Awards Awards for a challenge on “An ideal system architecture for HPC and AI converged requirements” |
Join us to this exciting Challenge and Win A Special Reward from Altair!
A Challenge on “An ideal system architecture for HPC and AI converged requirements”
The adoption of machine learning took skyrocket in the past few years. This has created a new computing demands and challenges. HPC simulations produce large amounts of data which ML can rely on and also ML applications can take benefit of HPC hardware and infrastructure. So, the HPC systems should be designed to serve both scientific and ML applications. Traditionally, HPC has larger compute, storage and fast network requirements. ML codes can run both on CPUs and accelerators (GPUs) but they are proved to perform faster and cost effective on GPUs.
Containers are playing a major role in delivering the ML software. The HPC scientific applications yet to adapt container-based software delivery model. So, there are mixed hardware and software requirements for HPC and ML to coexist. Many sites have proven that it is possible to have a converged infrastructure, but everyone have their own challenges.
This competition is about putting your ideas together for an ideal unified HPC and AI system.
Submission requirements :