NVIDIA's New Tesla GPU Accelerators to Boost Machine Learning

By CIOReview | Friday, November 20, 2015

FREMONT, CA; NVIDIA will be releasing the Tesla GPU Accelerators for servers to accelerate their high performance computing and machine learning. The two new hyperscale accelerators are Tesla M40 GPU and Tesla M4 GPU. The varied features of these accelerators are likely to support system management tools and standard applications of industries.

Tesla M40 GPU is the fastest deep learning accelerator. The major features of Tesla M40 GPU accelerator includes: Reduced Machine learning time by 8X faster than CPU, high performance and reliability. This enables the researchers to innovate, design deep new neural networks to enhance Artificial intelligence, so that it saves the amount of time needed for machine learning and improvises overall efficiency.

Tesla M4 accelerator is a low power accelerator mainly designed to drive machine learning in hyperscale data centers. The features include: optimized web service applications like video transcoding up to 5X faster, power efficient and small form factor. It optimizes video transcoding, image processing and machine learning interface and aims to serve the hyperscale environment.

NVIDIA Hyperscale Suite is a rich suite of software, optimized for machine learning and video processing. This serves both developers and data center managers mainly for web services deployments. The web services applications include: cuDNN, GPU, NVIDIA GPU REST Engine, NVIDIA Image Compute Engine.

Moreover, Mesosphere is collaborating with NVIDIA, to add support for GPU technology to Apache Mesos and the Mesosphere Datacenter Operating System (DCOS), in an idea to serve web –services companies to further work on the next-generation applications.