Vertiv Says 2.3MW CDU Meets NVIDIA DSX Ready Requirements
TECHWORLD ·
✦ AI Summary
Vertiv said on the 8th that it is strengthening its push into the cooling market for NVIDIA AI data center infrastructure with a 2.3MW cooling distribution unit (CDU).
It said its 2.3MW Chilled CDU model meets the requirements for "NVIDIA DSX Ready CDU," and that it is the first such case in the CDU lineup.
Vertiv plans to use this as an opportunity to expand its liquid cooling portfolio for a range of AI computing environments, including edge, colocation, and hyperscale.
Vertiv said on the 8th that it is stepping up its push into the cooling market for NVIDIA AI data center infrastructure with a 2.3MW cooling distribution unit (CDU). Vertiv said its 2.3MW Chilled CDU model meets the requirements for "NVIDIA DSX Ready CDU."
The model is the first in Vertiv's CDU lineup to meet the NVIDIA DSX Ready requirements. Vertiv has become one of the early suppliers offering CDU solutions that meet the requirements of the NVIDIA DSX Ready program.
The NVIDIA DSX Ready program helps AI factory builders identify power and cooling products that meet NVIDIA DSX-related requirements. The program also serves as a set of evaluation criteria in the process of selecting cooling products.
AI computing environments are becoming more dense. As a result, design requirements for power and cooling infrastructure are shifting from individual components to integrated systems.
Scott Armour, Vertiv's chief product and technology officer (CPTO), said that as the scale of AI builds expands, an integrated approach across data center infrastructure is needed. He said it is difficult to design and decide on computing, power, and cooling separately in a data center.
Scott Armour said the constraints of a single system can limit the performance and production capacity of the entire facility. He added that NVIDIA DSX Ready capabilities help evaluate core infrastructure from an integrated design perspective using consistent criteria.
Vertiv's 2.3MW CDU model is part of its liquid-to-liquid CDU portfolio. The model is designed to support direct-to-chip cooling and rear-door heat exchanger applications, and can be installed at the end of a row or at the edge of a data hall.
A CDU manages the flow rate and temperature of coolant to remove heat from high-density computing systems. In AI servers and other systems with high compute and power density, stable thermal management is considered a key infrastructure element.
Vertiv plans to expand its liquid cooling portfolio for a range of AI computing environments, including edge, colocation, and hyperscale, by meeting the DSX Ready requirements.
An AI factory is a data center designed for large-scale development and operation of AI models and applications. As computing density and infrastructure scale increase, the range of conditions under which an AI factory can be built is affected by power and cooling, water supply, site, and grid constraints.
The NVIDIA DSX AI Factory Platform integrates computing, networking, power, cooling, facilities, and software. The platform supports partners in designing and operating AI factory systems, and its design and operation take available power and cooling, water supply, and grid constraints into account.
Vertiv said meeting the 2.3MW CDU's DSX Ready requirements is an opportunity to strengthen its thermal management components, engineering expertise, and lifecycle support capabilities for AI environments. Scott Armour, Vertiv CPTO, said the 2.3MW CDU model's meeting of NVIDIA DSX Ready requirements demonstrates the company's efforts to continue providing thermal management components, engineering expertise, and lifecycle support for customers planning and operating increasingly complex AI environments.
However, meeting NVIDIA DSX Ready requirements does not replace the engineering needed for a specific site or system configuration. Customers need to separately review how the product should be applied to their facility design, system configuration, and operational requirements.
NVIDIA DSX Ready is a program that verifies whether individual products meet NVIDIA DSX-related requirements. In other words, its purpose is to confirm whether a given product meets the relevant requirements.
CDU suppliers make that determination using the NVIDIA CDU self-qualification suite. The NVIDIA CDU self-qualification suite is used to determine whether an individual product meets the relevant functional requirements.
Source: TECHWORLD · Park Gyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407921
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.