Whether we are engineering systems to power mobile devices or creating architectures that support high-performance supercomputers, improving energy efficiency is a principal goal in each step of our research, development, and design processes.
Parallel processing consumes far less power than equivalent computational forms. On a per-instruction basis, GPUs are dramatically more power efficient than CPUs, which have traditionally handled most instructional processing. The top 10 greenest supercomputers in the world are powered by NVIDIA technology, according to the latest Green 500 list.
Our highly efficient products and technologies include:
- NVIDIA Tesla solutions for high-performance computing: The Tesla K40 GPU accelerator outperforms CPUs by up to 10x. Scientists can run the most demanding scientific models on our GPU accelerators. The world's greenest supercomputer, the Tsubame-KFC system at the Tokyo Institute of Technology, was built using Tesla technology, and it performs at a whopping 4.5 gigaflops per watt.
- NVIDIA CUDA parallel processing architecture: CUDA is a parallel computing platform and computing model that enables compute-intensive calculations to be executed on lower cost, power-efficient GPUs. Use of the platform is gaining momentum — 35 institutions from 11 countries opened CUDA Research Centers and CUDA Teaching Centers in 2013, bringing the total to 273 in 42 nations. Learn more about GPU computing.
- NVIDIA Optimus technology: Optimus maximizes energy conservation and battery life in notebooks by automatically shutting off the GPU when it is not needed. Built using CUDA technology, Optimus seamlessly optimizes the user's computing experience while managing power usage behind the scenes.
- NVIDIA Tegra mobile processors: Tegra 4-Plus-1 technology delivers record levels of power and battery life. This variable symmetric multi-processing architecture enables four performance cores to be used for maximum bursts when needed, with each core independently and automatically enabled and disabled based on workload.
With the introduction of our Kepler architecture, NVIDIA's goal was to continue to push the limits of energy efficiency, based on performance per watt. In 2013, we extended Kepler's efficient architecture to mobile, enabling Kepler-powered tablet to use less than one-third the power while providing the same performance.
But we never rest on our laurels. In a massive leap forward for gamers, we introduced in February 2014 the first GPUs based on our new Maxwell™ graphics architecture. Maxwell doubles performance while using half the power of the GTX 550 Ti, which was built using Fermi architecture (which preceded Kepler). A recent white paper explains specific differences between the previous and current architectures and the innovative engineering that is improving energy efficiency.
NVIDIA Tegra K1 harnesses ARM's advanced CPU plus a third-generation battery-saver core, to deliver record levels of performance and battery life.
The top 10 world's most energy-efficient supercomputers are all powered by NVIDIA Tesla GPUs. Read more >