Accelerating scientific discovery, visualizing big data for insights, and providing smart services to consumers are everyday challenges for researchers and engineers. Solving these challenges takes increasingly complex and precise simulations, the processing of tremendous amounts of data, or training sophisticated deep learning networks. These workloads also require accelerating data centers to meet the growing demand for exponential computing.
NVIDIA Tesla is the world’s leading platform for accelerated data centers, deployed by some of the world’s largest supercomputing centers and enterprises. It combines GPU accelerators, accelerated computing systems, interconnect technologies, development tools, and applications to enable faster scientific discoveries and big data insights.
At the heart of the NVIDIA Tesla platform are the massively parallel GPU accelerators that provide dramatically higher throughput for compute-intensive workloads—without increasing the power budget and physical footprint of data centers.
Tesla® K80
Faster results and insights
The NVIDIA® Tesla® K80 is the world’s most powerful accelerator built for high-performance computing and machine learning applications. It delivers a 10x speed-up compared to the latest CPUs, and up to 4x acceleration over previous Tesla GPUs.
Take advantage of the powerful new features
Dual GPU Accelerator
With two GPUs delivering total memory bandwidth of 480 GB/s. data-intensive applications will experience the best performance on the K80.
GPU Boost
Realize more than 40% higher application performance by intelligently scaling clocks to maximum setting. Each application runs at peak performance while remaining within the power and thermal envelope.
24 GB of GPU Memory
Tesla K80 is the first accelerator to offer 24 GB of high-speed GDDR5 memory. That’s four times larger than the Tesla M2090 and K20, and big enough to accelerate the most data-intensive applications with less resources.
2x Shared Memory
Tesla K80 has double the on-chip memory registers, accelerating applications by up to 30% faster without any change to the code by enabling more threads to run concurrently.
Download Tesla® K80 datasheet >
Tesla® M4
The world’s first accelerator for the hyperscale data center
Exploding volumes of user-generated data are redefining what’s required for hyperscale data centers. Today’s cloud applications harness valuable data to deliver smarter, real-time experiences using modern video and image processing and deep learning techniques. These applications can benefit greatly from GPU acceleration in the data center.
The NVIDIA Tesla M4 is the world’s first accelerator designed for hyperscale servers, enabling customers to keep up with ever-growing amount of data. It’s engineered to accelerate application throughput in a small, low-power design, slashing data center costs by half and deliver up to 7x more power-efficient processing than CPUs for deep learning inference at 20 images/sec/watt and video workloads.
Download Tesla® m4 datasheet >
Tesla® M40
Power your data center with the world’s fastest deep learning training accelerator
Deep learning is redefining what’s possible, from image recognition and natural language processing to neural machine translation and image classification. From early-stage startups to large web service providers, deep learning has become the fundamental building block in delivering amazing solutions for end users.
Deep learning models typically take days to weeks to train, forcing scientists to make compromises between accuracy and time to deployment. The NVIDIA Tesla M40 GPU accelerator, based on the ultra-efficient NVIDIA MaxwellTM architecture, is designed to deliver the highest single precision performance. Together with its high memory density, this makes the Tesla M40 the world’s fastest accelerator for deep learning training.
Running Caffe and Torch on the Tesla M40 delivers the same model within hours versus days on CPU-based compute systems:
Deep Learning Ecosystem Built for Tesla Platform
The Tesla M40 accelerator provides a powerful foundation for customer to leverage best-in-class software and solutions for deep learning. NVIDIA cuDNN, DIGITS™ and various deep learning frameworks are optimized for the NVIDIA Maxwell™ architecture and Tesla M40 to power the next generation machine learning applications.