NVIDIA to launch A100 and H100 liquid-cooled PCIe accelerators

NVIDIA to launch A100 and H100 liquid-cooled PCIe accelerators
Written by admin_3fxxacau

Among NVIDIA’s slate of announcements tonight at Computex 2022, the company revealed that it’s gearing up to launch liquid-cooled versions of its high-end PCIe accelerator cards. Offered as an alternative to traditional dual-slot air-cooled cards, liquid-cooled cards come in a single, more compact slot for improved cooling and density. The liquid-cooled A100 will be available in the third quarter, and a liquid-cooled H100 will be available early next year.

While liquid cooling is far from new to the data center, it’s typically reserved for more custom hardware with extreme cooling and/or density requirements, such as the next generation of high-end NVIDIA H100 servers ( SMX). PCIe servers, on the other hand, are all about standardization and compatibility. Which, for video cards/server accelerators, means dual-slot cards designed for use with forced air cooling in a server chassis. This serves the market segment well, but the 300-350 watt TDPs of these cards mean they can’t get any thinner and can still be effectively air-cooled, which in turn creates a limitation of 4 cards for standard rackmount systems.

But times are changing and liquid cooling is being implemented in data centers with larger capacities, both to keep up with the cooling of ever hotter hardware and to improve the overall energy efficiency of the data center. To that end, NVIDIA will release liquid-cooled versions of its PCIe A100 and H100 cards to give data center customers an easy, officially supported way to install liquid-cooled PCIe accelerators in their installations.

The boards (shown above) are essentially an A100/H100 reference with the traditional dual-slot heatsink replaced with a single-slot full-coverage water block. Designed for integration by server vendors, they use an open loop design intended for use as part of a larger liquid cooling setup.

But apart from changing the cooling system, the specifications of the cards remain unchanged. NVIDIA doesn’t increase TDPs or clock speeds on these cards, so their performance should be identical to traditional air-cooled cards (as long as they’re not thermally throttled, of course). Simply put, these new cards use liquid cooling to improve power efficiency and density, rather than performance.

The first card to be released will be the liquid-cooled version of the 80GB PCIe A100 accelerator. This will be available to customers in the third quarter of this year. Meanwhile, a liquid-cooled version of the H100 PCIe is also in development, and NVIDIA expects it to be available in early 2023.

In the meantime, NVIDIA has been working with Equinix to qualify the liquid-cooled A100 within their data centers, as well as to get a sense of the real power savings of the new hardware. Interestingly, NVIDIA reports a significant reduction in overall data center power consumption with the switch to liquid cooling – a 2000 server configuration (4000 A100 cards) saw its total power requirements drop by 28%. According to NVIDIA, this comes from a combination of overall power savings in the data center from the switch, including everything from improved video card power efficiency to lower temperatures, to reduction of cooling water energy requirements compared to the use of large air coolers. All of this underscores why NVIDIA is promoting liquid-cooled hardware as an energy efficiency boost for data center operators looking to reduce power consumption.

And while this first generation of liquid-cooled hardware is all about efficiency, according to NVIDIA, that won’t always be the case. For future generations of cards, the company will also be looking at liquid cooling to improve performance at current power levels, likely by reinvesting data center-scale gains into higher TDPs for the cards.

Finally, while most of NVIDIA’s announcement today (as well as the case study) focuses on PCIe cards, NVIDIA also reveals that they’ve also been working on official liquid-cooled designs for their HGX systems, which are used to house the company’s most powerful SMX cards. A liquid-cooled HGX A100 is already shipping, and a liquid-cooled HGX H100 is slated for release in Q4.

#NVIDIA #launch #A100 #H100 #liquidcooled #PCIe #accelerators

About the author


Leave a Comment