CMKL University to Accelerate AI Research in Thailand with NVIDIA DGX POD

2020-10-17 12:28:59 PHT

Deploys advanced AI accelerated computing infrastructure

NVIDIA today announced another large deployment in Southeast Asia of the NVIDIA DGX POD AI accelerated computing infrastructure to drive artificial intelligence (AI) research at CMKL University in Thailand. The cluster will be the central computing node that connects to research and university nodes across the country.

As part of Thailand’s AI research infrastructure, CMKL is setting up the AI computing cluster to provide capabilities for data exchange, and an AI analytics platform to support competency building in the new normal economy. This will help drive research in various domains such as food and agriculture, healthcare and smart city.

“The platform that we are developing will allow researchers to store and manage their datasets with ease. They will be able to exchange their data with other researchers, and utilise the cloud HPC infrastructure to run machine learning codes and models at lightning speeds,” said Akkarit Sangpetch, CMKM Program Director (Thailand), CMKL University.

Established as a collaboration between Carnegie Mellon University (CMU) and King Mongkut's Institute of Technology Ladkrabang (KMITL), the university provides cutting-edge engineering research and education in Southeast Asia. By bridging world-class partnerships with local context, CMKL makes technologies accessible to society and creates innovations that will benefit Thailand and Southeast Asia. Applying CMU’s globally acclaimed research and education programmes within a regional context helps it tackle challenges that will drive future development of Thailand and the region.

Powering AI research projects

Modelled after NVIDIA DGX POD design, CMKL University’s new infrastructure consists of six NVIDIA DGX A100 AI systems delivering 30 petaflops of AI processing power, NVIDIA Cumulus networking software, NVIDIA Mellanox Spectrum 100GbE and Quantum 200Gbps InfiniBand smart switches that interconnect the processing nodes, and 2.5 petabytes of DataDirect Networks (DDN) storage.

When operational in early November, the NVIDIA DGX POD will run PyTorch for image processing, Kaldi for voice recognition, and NVIDIA Metropolis with the DeepStream SDK for smart city video analytics.

The system will deliver up to a 20x performance boost, which is greatly beneficial for AI projects such as automated speech recognition research that needs to collect and process 1,000 hours of voice input, and a bottle recycling model that contains 600,000 high-resolution images.

“This platform will accelerate AI work in various research and development fields to create substantial positive impacts on society. Examples are an increase in quality of life in cities through better management and logistics; better optimization of circular infrastructure, consumer insights, and biomedical research; and an increase in the quality and quantity of crop yields countrywide,” said Sangpetch.

Backing the project is Thailand’s Office of National Higher Education Science Research and Innovation Policy Council (NXPO), an autonomous public agency affiliated to the Ministry of Higher Education, Science, Research and Innovation.
“As NXPO is committed to increasing Thailand’s growth, level of competitiveness, and enhance its socio-economic sustainability through technology and innovation strategic plan. We are pleased to have PMU-C (Program Management Unit for Thailand’s Competitiveness) involved in providing the funding and support for this world-class research and initiative. Together with CMKL University and the new AI Infrastructure, the office looks forward to developing resilient economies in the age of AI,” said NXPO.

CMKL, which is currently running inference at the edge, is exploring using the NVIDIA Jetson AGX Xavier and NVIDIA Jetson Nano. It is also considering an extended cloud-edge processing model to utilise the inference capabilities both from the cluster as well as edge devices.

“NVIDIA DGX POD delivers groundbreaking performance that is designed to accelerate diverse AI training, inference and data science workloads. It will help accelerate CMKL University’s research to develop solutions that will benefit not just Thailand but also Southeast Asia and the world,” said Dennis Ang, Director, Enterprise Business, SEA and ANZ Region at NVIDIA.