NVIDIA adds liquid-cooled GPUs for sustainable and efficient computing

In response to the call for high-performance green data centers, NVIDIA announced the A100 PCIe liquid-cooled GPU, the first in a series of liquid-cooled GPUs for flagship servers. NVIDIA made this announcement during Computex 2022 in Taiwan.

In response to the call for high-performance green data centers, NVIDIA announced the A100 PCIe liquid-cooled GPU, the first in a series of liquid-cooled GPUs for flagship servers. NVIDIA made this announcement during Computex 2022 in Taiwan.

As part of a growing movement to build data centers that deliver high performance and energy efficiency, Zach Smith, Head of Advanced Infrastructure at Equinix, is committed to doing its part to be climate neutral. Zac runs more than 240 data centers for Equinix, a global service provider, so he knows the importance of reducing the carbon footprint of climate change.

Zack recounted that the 10,000 customers he supports rely on his team to deliver more data, more intelligence, often using artificial intelligence, and those customers want to deliver it more sustainably.

Marking progress in efficiency

As of April, Equinix has issued $4.9 billion in green bonds. These are investment-grade tools that Equinix will apply to reduce environmental impact by improving energy efficiency (PUE), an industry measure of how much energy a data center uses goes directly to computing tasks.

Data center operators are trying to bring this ratio closer to the ideal 1.0 PUE. Equinix facilities have an average PUE of 1.48 today, with their best new data centers coming in at less than 1.2.

Equinix is ​​making steady progress in the energy efficiency of its data centers as measured by the CPU (internal).

In January, Equinix opened a dedicated facility to track progress in energy efficiency. Part of this work focuses on liquid cooling. Born in the age of mainframes and maturing into the age of artificial intelligence, liquid cooling is now widely used within the world’s fastest supercomputers in a modern form called direct-chip cooling. Liquid cooling is the next step in accelerated computing for NVIDIA air-cooled GPUs that deliver up to 20 times better power efficiency on AI inference and high-performance computing functions than CPUs.

If you convert all AI-only CPU and HPC servers around the world to GPU-accelerated systems, you can save 11 trillion watt-hours of energy annually. This is equivalent to the energy use of more than 1.5 million homes per year.

NVIDIA is adding to its sustainability efforts with the launch of the first data center PCIe GPU using direct-chip cooling. Equinix qualifies the liquid-cooled NVIDIA A100 80GB PCIe GPU for use in its data centers as part of a comprehensive approach to sustainable cooling and heat capture. GPUs are sampling now and will be generally available this summer.

Save water and energy

According to Zac, the NVIDIA PCIe GPU is the first liquid-cooled GPU to be offered to the Equinix Lab, and Equinix customers are looking for sustainable ways to harness AI.

Data center operators aim to eliminate coolers that evaporate millions of gallons of water annually to cool the air inside data centers. Liquid cooling promises systems that recycle small amounts of liquids into closed systems that focus on major hotspots.

In separate tests, Equinix and NVIDIA found that a data center that uses liquid cooling can run the same workloads as an air-cooled facility while using 30 percent less energy. NVIDIA estimates that a liquid-cooled data center can reach 1.15 PUE, which is well below the 1.6 for its air-cooled cousin. Liquid-cooled data centers can also aggregate twice as much computing as the same space. That’s because A100 GPUs use only one PCIe slot, while air-cooled A100 GPUs fill two slots.

At least a dozen system makers plan to integrate these GPUs into their offerings later this year. They include ASUS, ASRockRack, Foxconn Industrial Internet, GIGABYTE, H3C, Inspur, Inventec, Nettrix, QCT, Supermicro, Wiwynn, and xFusion. NVIDIA also featured HGX reference designs and a fluid-assembled H100 GPU.

global trend

Regulations setting energy efficiency standards are pending in Asia, Europe and the United States. This motivates banks and other large data center operators to evaluate liquid cooling. Technology is not limited to data centers. Vehicles and other systems need them to cool high-performance systems built into confined spaces.

The path to sustainability

When asked, Zack explained that this is the beginning of the debut journey for the mainstream liquid-cooled accelerators. NVIDIA plans to follow up the A100 PCIe card with next year’s release with an H100 Tensor Core GPU based on the NVIDIA Hopper architecture and plans to support liquid cooling in high-performance data center GPUs and NVIDIA HGX platforms.

For rapid adoption, current liquid-cooled GPUs offer the same performance for less power. NVIDIA expects that in the future these cards will provide an option to get more performance for the same power; Something users say they want.

Learn more about what’s new A100 PCIe Liquid-Cooled GPUs.

Engage with StorageReview

the news | Youtube | audio notation Itunes/spotify | Instagram | Twitter | Facebook | tik tok | rss feed