Nvidia Launches Pascal, A GPU For The Cloud
March 27, 2014
Grazed from Cruxialcio. Author: Antone Gonsalves.
Nvidia has introduced its next-generation GPU platform, which is expected to significantly increase the speed of high-performance systems used in cloud computing and scientific research. Called Pascal, named after 17th century mathematician Blaise Pascal, the upcoming platform has a couple of proprietary technologies that move data much faster between the GPU and memory chips and the computer’s main processor, called the CPU.
GPUs are best known for rendering graphics in online gaming or videogames. However, the specialized design of the chips also makes them much better than CPUs at processing large blocks of data found in cloud-computing, machine learning, analytics and scientific research…
With Pascal, Nvidia is attacking the bottlenecks that exist when moving data onto and off the GPU. One potential problem area is in moving data between the GPU and its private memory chip, called the frame buffer or video memory, often located next to the GPU…
Read more from the source @ http://www.cruxialcio.com/nvidia-launches-pascal-gpu-cloud-5754
Subscribe to the CloudCow bi-monthly newsletter @ http://eepurl.com/smZeb