- NVIDIA GRID
- 3D Vision
- About NVIDIA
Disease diagnosis, environmental forecasting, surveillance for national security purposes, the search for clues along the far edge of the known universe — each is being addressed with tools accelerated through NVIDIA technology. Facilitating these breakthroughs is our core offering, the graphics processing unit, or GPU.
NVIDIA’s GPU works in tandem with a central processing unit (CPU), and is a specialized tool that breaks down complex computing problems into a great many smaller tasks that run concurrently. Our parallel processors accelerate a CPU by 10-100 times, enabling hundreds of billions of operations to occur each second. They turbocharge a system, rendering it more powerful and more efficient than a CPU alone.
Parallel processing is helping researchers make quantum leaps in progress in fields ranging from climate modeling to medical tomography, yielding advances that were previously impossible due to technology limitations. Problems that once took days to solve now can be addressed in minutes. Computations that previously required large, costly, energy intensive CPU clusters can now be accomplished more efficiently and less expensively by GPUs.
These advances are possible not only because we have invested billions of dollars in R&D, but also because we develop our products mindful of our ecosystem—our suppliers, partners and our end-customers, who are addressing enormously challenging problems.
In addition, we are engaged in efforts to reduce the power footprint of the GPU, focusing on improving our processors’ performance per watt of consumed power. We have made meaningful contributions through the deployment of Optimus, which intelligently optimizes a notebook computer’s battery life byshutting down the GPU when not in use—with no interference in usability or performance. Optimus reduces the GPU’s load on the battery and helps decrease the number of times it must be recharged, allowing for longer battery life. In addition, our Tegra system-on-a-chip, which is the size of a thumbnail, comprises eight separate processors that are activated only when required, enabling it to sip power. Designed for use in portable devices, it uses one-hundredth of the power traditionally used by desktop PC processors and one-tenth of the power of notebook processors.