Image source: Nvidia.com
We have been hearing about artificial intelligence every single day. People are talking about it on TED talk, podcast and articles.
Only time will tell how AI will impact our future (or present) but one thing we do know is that AI requires massive parallel computing power because it processes the humongous chunk of data. This application is more suited to GPU than general purpose CPU alone. GPUs are extremely adept at performing specific computation effectively and at lightning speed.
Recently, Nvidia announced a new GPU architecture called Volta, which contains 640 tensor cores capable of delivering 100 teraflops per second (TFLOPS) of deep learning performance. These tensor cores provide a specialised math cores that work in conjunction with the standard GPU CUDA cores to add additional processing for deep learning environments. Volta also includes over 5,000 GPU CUDA cores, 300 GB of system communications bandwidth through six high-speed NV Link interconnects, and 16 GB of the second generation high-bandwidth memory (HBM2) on TSMC’s new 12FFN manufacturing process technology.
It’s not only software tech giants like Microsoft, Google or Facebook who are targeting AI, but also semiconductor industry is keeping an eye on the emerging field of artificial intelligence and machine learning, and are building processors based on different architecture.
Click here to know more details.
The Future Chip 