Pages

Saturday, 3 November 2018

Taiwan's Chinese Medical University Hospital gets NVIDIA DGX-2 AI supercomputer

The Chinese Medical University Hospital (CMUH) in Taiwan is the first healthcare provider in Asia to deploy and operate the NVIDIA DGX-2 artificial intelligence (AI) supercomputer.

The DGX-2 system — provided through Ever Fortune.AI, a professional services partner — delivers new efficiencies and possibilities through GPU-accelerated applications. It is the first AI system to deliver 2 petaFLOPS of computational performance powered by the integration of 16 NVIDIA Tesla V100 Tensor Core GPUs using the AI network fabric, NVSwitch.

Taiwan is one of the most promising markets for AI-powered healthcare. The popular national health insurance system and more than two decades of electronic medical records provide Taiwan with massive amounts of coded data, images and video for big data analysis. With the DGX-2, researchers can leverage these enormous datasets and the most complex AI models at unprecedented speed.

Dr Fuu Jen Tsai, VP, CMUH and an expert in genetics and metabolics said, “The complexities of healthcare have resulted in slower AI adoption. While AI is becoming ubiquitous, we have yet to realise the game-changing clinical, operational, and financial opportunities that await us in healthcare. With Ever Fortune.AI’s support, AI has begun making progress to solve process inefficiencies, relieve tedious and expensive procedures, and guard against human error. It promises to usher in a new era of patient care.”

“With the help of the NVIDIA DGX-2 system, we are now able to accelerate our solutions by processing larger amounts of data and building more complex models,” said Dr Eddie Huang, Director of Medical Diagnostic Center, CMUH research team, and GM at Ever Fortune.AI. “We are
gaining momentum in Taiwan’s medical history to realise the era of smart healthcare.”

Beginning with building image-labelling and data-collection systems to standardise data for AI training, Ever Fortune.AI has been working closely with many medical practitioners at CMU. The team has launched several solutions, including a breast cancer classification system, a liver cancer evaluation and management system, and a bone age assessment system, which are now adopted by 16 hospitals across Taiwan.

“Global efforts to improve healthcare outcomes require innovation and NVIDIA DGX-2 is the world’s most powerful AI system for the most complex AI healthcare challenges,” said Eunice Chiu, VP of Sales and GM in Taiwan at NVIDIA.

“With CMUH, Ever Fortune.AI is embedding AI into clinical workflows, which will yield profound results for clinicians and their patients in Taiwan and potentially around the world.”

The news follows the certification of the NVIDIA DGX-1 for Red Hat Enterprise Linux (RHEL).
The combination enables enterprises to easily plug in and power up the AI supercomputer with a familiar operating system. According to NVIDIA, 90% of Fortune 500 companies use Red Hat Enterprise Linux in their operations.

This certification provides a foundation for the rest of the Red Hat portfolio, including Red Hat OpenShift Container Platform, to be deployed and jointly supported on NVIDIA’s AI supercomputers.

Beyond Red Hat Enterprise Linux certified and supported on NVIDIA DGX systems, the companies intend to collaborate on broader open source initiatives, including:
  • NVIDIA GPU Cloud (NGC) containers on Red Hat OpenShift

    Red Hat and NVIDIA plan to deliver NGC containers that provide users with GPU-optimised software tools for AI and HPC based on Red Hat technologies, enabling mutual customers to take full advantage of NVIDIA GPUs.

  • Heterogeneous memory management (HMM)

    Red Hat and NVIDIA plan to continue upstream development of the HMM feature. This kernel feature allows devices to access and mirror the content of a system’s memory into their own, enabling significant performance improvements for applications using GPUs.
NVIDIA DGX systems are built on the NVIDIA Volta GPU platform. They combine GPU-optimised software and simplified management.

Hashtags: #DGX2, #DGX1

No comments:

Post a Comment