ASRock Unveils”World’s Smallest” Server Rack, Featuring NVIDIA GH200 Superchips

ASRock has pushed a new server rack, which is said to be the "smallest" to integrate NVIDIA's cutting-edge GH200 Superchips.

With rapid advancements in AI and data centers, firms have stepped up in the race to develop new solutions that provide the required computing power much more effectively. Similarly, ASRock Rack, one of the ASRock divisions, has unveiled the new ASRock Rack MECAI-GH200, which is advertised to be the smallest server rack to house Team Green's Grace Hopper GH200 module, and is undoubtedly a technological marvel when it comes to providing the necessary power in a confined environment.

As ASRock Rack, we aspire to bring the benefits of AI to everywhere. To achieve this goal, we are committed to deliver the most dependable server hardware solutions in various form factors for different scenarios. We are glad about the opportunity to showcase the MECAI-GH200 as one example of our ambition for 2024 here at the Mobile World Congress, enabling AI on the edge.

- Hunter Chen, Vice President at ASRock Rack

For a quick rundown on the specification of the Grace Hopper Superchip, it comes in two different variants, one as the Grace Hopper Superchip with the H200 GPU and a single Grace chip with 72 Neoverse V2 cores with HBM3 memory. At the same time, the second model is the Grace Superchip, which features two Grace CPUs, each with 72 cores for a total of 144 cores with LPDDR5x memory. To create a seamless integration amongst the onboard components, NVIDIA utilizes the NVLink interconnect technology, which provides the best performance out of what is available.

In addition to the ASRock solution, NVIDIA has also announced several Grace Hopper GH200 Super-chip products which include:

ASRock RACK’s MECAI, measuring 450 x 445 x 87mm, accelerates AI and 5G services in constrained spaces at the edge of telco networks.

ASUS’s MGX server, the ESC NM2N-E1, slides into a rack that holds up to 32 GH200 processors and supports air- and water-cooled nodes.

Foxconn provides a suite of MGX systems, including a 4U model that accommodates up to eight NVIDIA H100 NVL PCIe Tensor Core GPUs.

GIGABYTE’s XH23-VG0-MGX can accommodate plenty of storage in its six 2.5-inch Gen5 NVMe hot-swappable bays and two M.2 slots.

Inventec’s systems can slot into 19- and 21-inch racks and use three different implementations of liquid cooling.

Lenovo supplies a range of 1U, 2U, and 4U MGX servers, including models that support direct liquid cooling.

Pegatron’s air-cooled AS201-1N0 server packs a BlueField-3 DPU for software-defined, hardware-accelerated networking.

QCT can stack 16 of its QuantaGrid D74S-IU systems, each with two GH200 Superchips, into a single QCT QoolRack.

Supermicro’s ARS-111GL-NHR with nine hot-swappable fans is part of a portfolio of air- and liquid-cooled GH200 and NVIDIA Grace CPU systems.

Wiwynn’s SV7200H, a 1U dual GH200 system, supports a BlueField-3 DPU and a liquid-cooling subsystem that can be remotely managed.

Wistron’s MGX servers are 4U GPU systems for AI inference and mixed workloads, supporting up to eight accelerators in one system.

The new and compact server rack by ASRock will catalyze AI developments in an edge environment, providing the best computing power available in the markets.