Nvidia hgx

Nvidia HGX-2 is a cloud server platform equipped with 16 GV100 chips through the use of NVSwitch interconnects, which work together offering half a terabyte of memory and two petaflops of computing power, spectacular figures only within reach of the green giant.

Nvidia HGX-2 is the new monster based on Volta architecture

According to a statement from Nvidia, the Nvidia HGX-2 can process 15, 500 images per second in the ResNet-50 training benchmark test, and can replace up to 300 CPU-based servers, which is a significant money saver. Server manufacturers like Lenovo and other manufacturers like Foxconn plan to launch the first HGX-2-based systems later this year. Enhancements to Nvidia's deep learning computing platform hardware and software have led to a 10-fold performance boost in deep learning workloads within six months.

We recommend you read our post on What graphics card do I buy? The best on the market 2018

A recent OpenAI study found that the computing power required to train well-known systems has doubled once every 3.5 months since 2012. This causes Nvidia's GPUs to be increasingly used for training and displaying untagged data like photos and videos.

“The scale of the CPU has slowed down at a time when computing demand is soaring. The Nvidia HGX-2 with Tensor Core provides the industry with a powerful and versatile computing platform that fuses HPC and AI to solve the world's great challenges. ”

The introduction of the Nvidia HGX-2 follows the launch of the HGX-1 last year, which runs on eight GPUs. The HGX-1 reference architecture has been used in GPU servers such as Facebook's Big Basin and Microsoft's Project Olympus that they desire.

Techpowerup font