To curb climate change, global companies are accelerating the construction of high-performance, energy-efficient data centers, and Zac Smith, head of Equinix’s edge infrastructure, is also involved.
The Equinix he works for is a global service provider with more than 240 data centers managed by it, committed to being the first in the industry to achieve climate neutrality.
"Ten thousand customers rely on Equinix to achieve climate neutrality. They need more data and higher levels of intelligence, which is usually dependent on AI, and they want to be able to achieve it in a sustainable way," Smith said. In the early 2000s, he was still studying for a graduate school at the Juilliard School and built a website for his fellow musicians in New York, thus forming an indissoluble bond with technology.
energy efficiency has steadily improved
As of April, Equinix has issued US$4.9 billion in green bonds. Equinix will use these investment tools to optimize power usage efficiency (PUE) in order to reduce environmental impact. PUE is an industry metric that measures how much energy a data center uses is directly used for computing tasks.
Data center operators are trying to reduce PUE to an ideal level close to 1.0. The Equinix facility currently has an average PUE of 1.48, while its new data center can have a minimum PUE of less than 1.2.
From the perspective of PUE, Equinix is steadily improving the energy efficiency of data centers (see illustration)
Equinix launched a new facility dedicated to improving energy efficiency in January, and went further in this direction. Some of the work focuses on liquid cooling technology.
Liquid cooling technology was born in the era of mainframes and is becoming increasingly mature in the era of AI. Today, liquid cooling technology has been widely used in high-speed supercomputers around the world in the form of direct-to-Chip cooling.
NVIDIA GPU has an energy efficiency of 20 times higher than that of CPUs in AI inference and high-performance computing, and accelerated computing will naturally adopt liquid cooling technology.
Improves efficiency through acceleration
If all CPU servers running AI and HPC in the world are switched to GPU acceleration systems, up to 11 trillion watt-hours of energy can be saved every year. The energy savings can be used for more than 1.5 million homes in one year.
Today, NVIDIA released the PCIe GPU, the first data center that uses Direct-to-Chip cooling technology, which contributed its own strength to sustainable development.
Equinix is verifying the application of the A100 80GB PCIe liquid-cooled GPU in its data centers, which is part of the company's comprehensive solution to achieve sustainable cooling and heat capture. The GPU has now entered trial phase and is expected to be officially released this summer.
Save water and electricity
"This is the first liquid-cooled GPU introduced in our lab, and we are excited because our customers are eager to leverage AI in a sustainable way," Smith said.
Data center operators aim to eliminate chillers used to cool gases inside data centers because they evaporate millions of gallons of water each year. With the help of liquid cooling technology, the system only needs to recycle a small amount of liquid in the closed system and can focus on the main heating points.
Smith said: "We will turn waste into treasure."
has the same performance and less power consumption
In separate tests, Equinix and NVIDIA both found that data center workloads with liquid cooling technology can be the same as air-cooled facilities, while consuming about 30% less energy. NVIDIA estimates that the PUE in liquid-cooled data centers may reach 1.15, which is much lower than the air-cooled PUE 1.6.
Under the same space, the liquid-cooled data center can achieve double the calculation amount. This is because the A100 GPU uses only one PCIe slot, while the air-cooled A100 GPU requires two PCIe slots.
NVIDIA has achieved power saving and density increase with liquid cooling technology
At least a dozen system manufacturers plan to use liquid cooling GPUs in their products later this year, including Asus, ASUS, ASRock Rack, Foxconn Industrial Internet, GIGABYTE, H3C, Inspur, Inventec, Nettrix, Vanda Technology (QCT), Supermicro, Wiwynn and xFusion.
Global Trends
In Asia, Europe and the United States, regulations on setting energy efficiency standards have not yet been determined. This has also pushed banks and other large data center operators to join the liquid cooling technology evaluation team.
The scope of use of liquid cooling technology is not limited to data centers. Automobile and other systems also need to use this technology to cool high-performance systems in closed spaces.
The Road to Sustainable Development
"We are about to embark on a new journey," Smith said of the mainstream accelerator that uses liquid cooling technology for its debut.
In fact, NVIDIA plans to launch next year's A100 PCIe card with the H100 Tensor Core GPU based on the NVIDIA Hopper architecture. In the near future, NVIDIA plans to apply liquid cooling technology to its own high-performance data center GPUs and NVIDIA HGX platforms.
To promote rapid adoption, the liquid-cooled GPU released today can maintain performance while reducing energy consumption. In the future, we expect these cards to show better performance and meet user needs under the conditions of using the same energy.
"It doesn't make sense to measure power just by reducing carbon emissions while improving performance is the direction we are working on," Smith said.