3 ways to advance sustainability in high performance computing

Finding the answer to the world’s most pressing issues rests on one crucial capability: high performance computing (HPC). With HPC, complex questions that have puzzled humankind for centuries are being unraveled at record speeds–such as unlocking mysteries of the universe, finding cures for diseases, sequencing DNA, and mitigating the impacts of climate change. 

The supercomputers that power HPC, however, require more and more energy to operate. For instance, the power consumption of the world’s fastest supercomputer rose from 7.9MW in 2012 to 29.9MW in 2022. It’s no wonder that–per a recent study commissioned by Dell Technologies, Intel and NVIDIA–HPC operators have elevated the importance of sustainability to the number two priority, even surpassing price. 

We applaud and support the efforts of HPC operators to improve sustainability. That means we must collectively and continuously work to manage HPC’s power requirements in areas where we can have a measurable impact. There are three ways to do this:

1. Maximize hardware energy efficiency 
2. Consolidate infrastructure
3. Use renewable energy sources

Deploying Energy-Efficient Hardware

The Hyperion Research study found three geo-specific motivations for the heightened prioritization of sustainability in HPC deployments. These motivations help shape geo-specific conversations about sustainability. As depicted in the figure below, 

• The Asia-Pacific region is driven by environmental, social, and governance (ESG) goals
• In Europe, organizations are concerned about rising, and thus unmanageable energy costs
• North America is motivated to do more with less

Universally, 60% of respondents in the study said sustainability initiatives will impact their on-premises HPC budgets. Interestingly, half of those respondents said sustainability initiatives will reduce on-premises purchases; the other half indicated that sustainability initiatives will increase them.

For respondents anticipating an HPC budget reduction, the goal is to reallocate dollars to sustainability initiatives. Dollars are moving to purchasing new energy-efficient hardware or devoting resources to optimization efforts or changing where HPC workloads are run.

Dell Technologies, NVIDIA, and Intel are supporting these organizational efforts. For instance, Dell Technologies, a committed steward of sustainability, has worked to decrease energy intensity across its entire portfolio, achieving a 76% reduction since 2013.

Consolidating Infrastructure

Beyond the individual hardware components, designing and deploying HPC infrastructure is a sophisticated undertaking. Ensuring infrastructure is consolidated and set up for optimal performance is a key lever to accelerating sustainability initiatives. At Durham University, this principle is fundamental to providing HPC resources. Since 2001, Durham University’s tier 1 national supercomputing facility has supported scientists and researchers around the world and is using HPC to build a digital simulation of the universe, starting with the Big Bang.

During an infrastructure upgrade, Dell Technologies helped Durham University optimize and modernize its infrastructure to achieve performance enhancements and improved sustainability. The results include 18X faster data backups, 72% less power, and a reduction of 60 tons of CO₂ per year.

Using Renewable Energy Sources

In addition to reducing the energy intensity of hardware and consolidating HPC infrastructure, sustainability initiatives are also advancing with the use of renewable energy sources. A great example of that is atNorth and BNP Paribas. When it came time to update its infrastructure, BNP Paribas, a leading bank in the European Union, turned to atNorth and Dell Technologies to help it expand responsibly and build a “future-proofed” HPC infrastructure.

BNP Paribas moved a portion of its data center operations to atNorth’s facility in Iceland, an economical, energy-efficient data center operating 100% on renewable energy. As a result, BNP Paribas lowered its total cost of ownership (TCO) with 50% less energy and 85% less CO₂ output. In addition, BNP Paribas increased power efficiency at higher compute density, achieving future-proofed, more environmentally-responsible HPC resources.

“By using only renewable energy sources and decreasing our carbon footprint by 85%, BNP Paribas is realizing its dual mission to reduce its environmental impact and better serve our customers,” shared Ricardo Jantarada, Global Head of Telecom & Datacenter at BNP Paribas. 

Advancing Environmentally Sustainable HPC

HPC holds great promise to unlock some of humankind’s most pressing, complex issues. At the same time, deploying HPC sustainably has risen to a top priority, as shown by research with organizations around the world. To advance sustainable HPC deployments requires action in three major areas: hardware energy efficiency, infrastructure consolidation, and renewable energy sources. Starting today helps accelerate sustainability initiatives to create a better tomorrow for all. 

Want to learn more about HPC’s amazing possibilities? View The Seven Wonders of the HPC World here.

Learn more about Dell Technologies’ environmentally sustainable HPC solutions here. 

______________

Sponsored by Dell Technologies, Intel and NVIDIA, the Hyperion Research survey had worldwide coverage with a strong focus on international sites and sustainability questions relating to HPC, artificial intelligence (AI), quantum computing, and the results of using HPC. Data was collected via direct interviews and surveys of HPC and Cloud Service Provider (CSP) data center managers, directors, or leads as well as scientists, researchers, and engineers.