Nvidia boosts AI in medical instruments with graphic compute power

Find the Latest SoC Solutions
for... Automotive... IoT... Security... Audio... Video

GPU Technology Conference, Munich: Nvidia is bringing artificial intelligence to medical instruments, with the Clara platform, based on the company's Xavier AI computing module and Turing GPUs introduced earlier this year, and the Clara software development kit.

electronicsweekly.com, Oct. 10, 2018 – 

It has been developed to address the increasing number of computing functions in today's medical healthcare, with CT and MRI scans increasing, while the number of radiologists has remained the same, meaning that the 20 seconds usually taken to look at an image has reduced to between three and five seconds, explained Kimberly Powell, vice president of healthcare, Nvidia. Medical data volume is measured in thousands of Gbytes, generated each second to be interpreted by doctors and scientists, requiring the computation abilities of fpgas, cpus and gpus.

Clara AGX distils these architectures into a single gpu-based architecture that is claimed to deliver the fastest AI inferencing on Nvidia Tensor Cores. Acceleration is using the Cuda accelerated computing platform (now on version six) and the company's RTX graphics.

 Back