May 11 2017
Vizua and Orange Silicon Valley are prototyping a new GPU-enabled Cloud. Based on Vizua's server computing patents and supercomputing platform from Orange Silicon Valley, the new Cloud will serve the emerging artificial intelligence (AI) and augmented reality (AR) applications in a wide range of industries.
Because of increasing data and sophisticated algorithms for processing and display, the biggest challenge today for AI and AR applications is computing power. Additionally, investing in dedicated resources can be costly -- not only for upfront acquisition, but also maintenance and depreciation. And, dedicated resources often are not accessible to every device, everywhere, and make it difficult to share data and experiences. The new cloud prototype from Vizua/Orange Silicon Valley works towards addressing all these problems.
The new cloud prototype provides high speed GPU enabled processing on-demand and is accessible to devices everywhere. A single server has supercomputing capability with 20 NVIDIA GPUs capable of supporting real time 3D rendering of 200 Microsoft HoloLens users.
This supercomputing platform with optimized software from Vizua3D can handle over 10 times more users per GPU than a regular virtual machine (VM). It can also process extremely large files by optimizing the distribution over GPUs. Terabytes of data can be managed and loaded in a few seconds, and it is possible to launch a cloud service by leveraging such supercomputing infrastructure powered by the latest NVIDIA Pascal GPU architecture. Furthermore, the cloud based solution is designed to address a wide spectrum of markets: radiology, architecture, design, retail, industry, archeology, and gaming, to name just a few.
In the medical field, TeraRecon, with its Within Image Analysis product (WIA™), is leveraging the supercomputing testbed infrastructure from Orange Silicon Valley to accelerate AI workloads. The prototype AI platform from Orange Silicon Valley allows WIA™ to boost performance for deep neural network based training, testing and application of the AI engine to improve clinical workflow as well as the diagnosis of patients.
"We are very pleased to partner with the Orange Silicon Valley team," says Sylvain Ordureau, co-founder and CEO of Vizua. "Orange Silicon Valley and Vizua are prototyping a new HPC Cloud GPU platform to optimize the cost, performance and number of users per GPU to enable modern, intelligent applications in a wide variety of markets."
"Deep Learning on hundreds of Tera Bytes of imaging data requires enormous computational capability and real time holographic rendering workloads with augmented cognitive capabilities are demanding when we scale to hundreds of users. Real-time 3D imaging combined with Artificial Intelligence powered inferencing capability will revolutionize medicine," says Georges Nahon CEO of Orange Silicon Valley.
Vizua will be at GTC 2017 San Jose (CA) May 7-11 to demonstrate the power of their technology on the supercomputing server prototype from Orange Silicon Valley. Booth visitors can experience the Microsoft HoloLens and view shared virtual models projected on real objects with very high precision and incredible image quality. AI medical applications showcasing 3D printed medical models and TeraRecon software (terarecon.com) will also be demonstrated.