#acceleratedcomputing

For GPU and CPU processors in accelerated systems, NVLink is a high-speed link that drives data and computations to useful outcomes. Once limited to high-performance computers in government research facilities, accelerated computing has become widely available.

AI supercomputers are being used by banks, automakers, factories, hospitals, merchants, and others to handle the increasing amounts of data that they must process and comprehend.

These strong, effective systems are computing superhighways. On their lightning-fast trip to actionable answers, they transport calculations and data via parallel routes.

The resources along the route are CPU and GPU processors, and their onramps are quick connections. NVLink is the industry standard for accelerated computing interconnects.

What is NVLink?

A reliable software protocol creates the high-speed link between GPUs and CPUs known as NVLink, which usually runs on many wire pairs printed on a computer board. It enables lightning-fast data transmission and reception across processors from shared memory pools.Image Credit To NVIDIA

At speeds of up to 900 gigabytes per second (GB/s), NVLink, which is now in its fourth iteration, links host and accelerated processors.

The bandwidth of PCIe Gen 5, the link found in traditional x86 servers, is more than seven times that amount. Additionally, because NVLink uses just 1.3 picojoules per bit for data transfers, it has five times the energy efficiency of PCIe Gen 5.

History of NVLink

With each successive NVIDIA GPU architecture, NVLink has improved in tandem since its initial release as a GPU connection with the NVIDIA P100 GPU.Image Credit To NVIDIA

When NVLink first connected the GPUs and CPUs in two of the most potent supercomputers in the world, Summit and Sierra, in 2018, it gained significant attention in the field of high performance computing.

Installed in Oak Ridge and Lawrence Livermore National Laboratories, the systems are advancing science in areas like drug development and catastrophe forecasting, among others.

Bandwidth Doubles, Then Grows Again

The third-generation NVLink, which came with a dozen interconnects in each NVIDIA A100 Tensor Core GPU, increased the maximum capacity per GPU to 600GB/s in 2020.

AI supercomputers in cloud computing services, business data centers, and HPC laboratories worldwide are powered by the A100.

One NVIDIA H100 Tensor Core GPU now contains eighteen fourth-generation NVLink interconnects. Additionally, the technology has assumed a new, strategic function that will allow for the world’s most sophisticated CPUs and accelerators.

A Chip-to-Chip Link

A board-level connection called NVIDIA NVLink-C2C is used to combine two processors onto a single device, forming a superchip. For instance, the NVIDIA Grace CPU Superchip, a processor designed to provide energy-efficient performance for cloud, corporate, and HPC customers, combines two CPU chips to produce 144 Arm Neoverse V2 cores.

A Grace CPU and a Hopper GPU are additionally joined via NVIDIA NVLink-C2C to form the Grace Hopper Superchip. It combines accelerated computation for the most demanding AI and HPC tasks on a single chip.

One of the first to employ Grace Hopper will be Alps, an AI supercomputer slated for the Swiss National Computing Center. The high-performance supercomputer will tackle huge research challenges in domains ranging from quantum chemistry to astrophysics when it goes online later this year.Image Credit To NVIDIA

Additionally, Grace and Grace Hopper are excellent in reducing energy consumption in demanding cloud computing tasks.

The Grace Hopper processor, for instance, is perfect for recommender systems. In order to provide billions of users with trillions of results every day, these internet economic engines require quick and effective access to large amounts of data.

A potent system-on-chip for automakers that incorporates NVIDIA Hopper, Grace, and Ada Lovelace processors and uses NVLink. NVIDIA DRIVE Thor is a vehicle computer that integrates cognitive features including entertainment, automatic driving, parking, digital instrument cluster, and more into a single architecture.

LEGO Links of Computing

NVLink functions similarly to the socket that is imprinted onto a LEGO component. It serves as the foundation for constructing supersystems to handle the most challenging AI and HPC tasks.

For instance, an NVIDIA DGX system’s eight GPUs’ NVLinks exchange quick, direct connections using NVSwitch chips. When combined, they allow for an NVLink network in which each server’s GPU functions as a single system.

DGX workstations itself may be stacked into modular units of 32 servers to provide even higher performance, forming a strong, effective computing cluster.

Using an NVLink network within the DGX and an NVIDIA Quantum-2 switched InfiniBand fabric between them, users may integrate a modular block of 32 DGX devices into a single AI supercomputer. An NVIDIA DGX H100 SuperPOD, for instance, has 256 H100 GPUs to provide up to an exaflop of the best AI performance.

Users may access cloud-based AI supercomputers, like the one Microsoft Azure is constructing with tens of thousands of A100 and H100 GPUs, to achieve even higher performance. Some of the biggest generative AI models in the world are trained using this service by organizations like OpenAI.

Additionally, it serves as another illustration of the potential of accelerated computing.

Artificial intelligence (AI) and accelerated computing are revolutionizing businesses by promoting energy efficiency and providing creative answers to pressing global environmental issues.

Naturally, AI systems use energy, but the real story here is how AI and faster computing are assisting in the energy conservation of other systems. Consider data centers as an example.

They serve as the foundation of artificial intelligence (AI), containing the robust systems that process the data required for AI to function in applications like chatbots, AI-powered search, and content creation. AI-specific centers make up a very small portion of the data centers’ global energy usage, which is around 2%.

All the same, AI’s true strength is its optimization capability

How? via the use of systems for accelerated computing that integrate CPUs and GPUs. GPUs are designed to do complicated calculations fast and effectively.

AI can now handle more complicated issues in part because of these advancements, which have led to enormous advances in energy efficiency over the last eight years.

What Is Accelerated Computing?

Accelerated computing is fundamentally about achieving more with less. It entails using specialized technology, like as GPUs, to do jobs more quickly and energy-efficiently. Businesses may expand their AI operations with the use of accelerated computing without using enormous amounts of energy.

AI’s potential to address some of the most pressing ecological issues of day depends on its energy efficiency.

AI in Action: Addressing Global Warming

AI is combating climate change in addition to saving electricity.

Artificial intelligence and faster Accelerated Computing in general are revolutionizing weather and climate modeling and simulation.

For example, governments and businesses are able to better plan for climate-related disasters like hurricanes and floods because to the increasing accuracy of AI-enhanced weather forecasting.

NIVIDIA ability to anticipate these catastrophes will help us plan for them better, which will save resource waste and harm.

The emergence of digital twins, or virtual representations of real settings, is another important trend.

Businesses can now optimize energy use in real time without having to make expensive physical modifications to the environment with these AI-powered simulations.

That may not seem like much, but applying it to numerous areas has huge impact.

In order to hasten the switch to clean energy, artificial intelligence is also essential in the development of novel materials for renewable energy technologies like solar panels and electric cars.

Does AI Improve Data Center Sustainability?

AI requires data centers and processing capacity as it grows. Yet, data centers don’t always need a lot of energy. They may even contribute to the sustainable solution.

One significant advancement is liquid cooling directly to the chip. Data centers can now cool their systems much more effectively than they could with more energy-intensive conventional air conditioning techniques with this technology.

“Direct-to-chip liquid cooling is the main focus of recommended design for the data centers for a new B200 chip. This approach reduces energy consumption by cooling data centers directly at the chip level, keeping them cold without using excessive power.

The future of data centers will rely on building them with energy efficiency from the bottom up as AI becomes more and more prevalent.

This entails using energy storage technologies, incorporating renewable energy sources, and keeping up with cooling technology innovation.

The goal is to construct environmentally friendly data centers that can meet the growing demand for processing power globally without increasing their carbon footprint.

AI’s Place in Creating a Sustainable Future

Artificial Intelligence is a catalyst for long-term innovation, not only a tool for system optimization.

AI is reducing waste and pollution by improving supply chain operations and energy grid efficiency.

AI “will be the best tool we’ve ever seen to help us achieve more sustainability and more sustainable outcomes.”