ASRock Rack sells the entire stack of servers a data-center could possibly want, and at Computex 2024, the company showed us their servers meant for AI GPUs. The 6U8M-GENOA2, as its name suggests, is a 6U server based on 2P AMD EPYC 9004 series "Genoa" processors in the SP5 package. You can configure it with even the variants of "Genoa" that come with 3D V-cache, for superior compute performance from the large cache. Each of the two SP5 sockets is wired to 12 DDR5 RDIMM slots, for a total of 24 memory channels. The server supports eight AMD Instinct MI300X or MI325X AI GPUs, which it wires out using Infinity Fabric links and PCIe Gen 5 x16 individually. A 3 kW 80 Plus Titanium PSU keeps the server fed. There are vacant Gen 5 x16 slots left even after connecting the GPUs, so you could give it a DPU-based 40 GbE NIC.

The 6U8X-EGS2 B100 is a 6U AI GPU server modeled along the 6U8M-GENOA2, with a couple of big changes. To begin with, the EPYC "Genoa" chips make way for a 2P Intel Xeon Socket E (LGA4677) CPU setup, for 2P Xeon 5 "Emerald Rapids" processors. Each socket is wired to 16 DDR5 DIMM slots (the processor itself has 8-channel DDR5, but this is a 2 DIMM-per-channel setup). The server integrates an NVIDIA NVSwitch that wires out NVLinks to eight NVIDIA B100 "Blackwell" AI GPUs. The server features eight HHHL PCIe Gen 5 x16, and five FHHL PCIe Gen 5 x16 connectors. There are vacant x16 slots for your DPU/NIC, you can even use an AIC NVIDIA BlueField card. The same 3 kW PSU as the "Genoa" system is also featured here.

Intel today launched its Arctic Sound M line of data-center GPUs. These are not positioned as HPC processors like the "Ponte Vecchio," but GPUs targeting cloud-compute providers, with their main applications being in the realm of visual processing, media, and AI inferencing. Their most interesting aspect has to be the silicon, which are the same 6 nm "ACM-G11" and "ACM-G10" chips powering the Arc "Alchemist" client graphics cards, based on the Xe-HPG architecture. Even more interesting is their typical board power values, ranging between 75 W to 150 W. The cards are built in the PCI-Express add-on card form-factor, with their cooling solutions optimized for rack airflow.

The marketing name for these cards is simply Intel Data Center GPU Flex, with two models being offered: The Data Center GPU Flex-140, and Flex-170. The Flex-170 is a full-sized add-on card based on the larger ACM-G10 silicon, which has 32 Xe Cores (4,096 unified shaders), whereas the Flex-140, interestingly, is a low-profile dual-GPU card with two smaller ACM-G11 chips that each has 8 Xe Cores (1,024 unified shaders). The two chips appear to be sharing a PCIe bridge chip in the renders. Both models come with four Xe Media Engines that pack AV1 encode hardware-acceleration, XMX AI acceleration, real-time ray tracing, and GDDR6 memory.

Super Micro Computer, Inc., a global leader in enterprise computing, storage, networking, and green computing technology, is announcing future Total IT Solutions for availability with Android Cloud Gaming and Media Processing & Delivery. These new solutions will incorporate the Intel Data Center GPU, codenamed Arctic Sound-M, and will be supported on several Supermicro servers. Supermicro solutions that will contain the Intel Data Center GPUs codenamed Arctic Sound-M, include the 4U 10x GPU server for transcoding and media delivery, the Supermicro BigTwin system with up to eight Intel Data Center GPUs, codenamed Arctic Sound-M in 2U for media processing applications, the Supermicro CloudDC server for edge AI inferencing, and the Supermicro 2U 2-Node server with three Intel Data Center GPUs, codenamed Arctic Sound-M per node, optimized for cloud gaming. Additional systems will be made available later this year.

"Supermicro will extend our media processing solutions by incorporating the Intel Data Center GPU," said Charles Liang, President, and CEO, Supermicro. "The new solutions will increase video stream rates and enable lower latency Android cloud gaming. As a result, Android cloud gaming performance and interactivity will increase dramatically with the Supermicro BigTwin systems, while media delivery and transcoding will show dramatic improvements with the new Intel Data Center GPUs. The solutions will expand our market-leading accelerated computing offerings, including everything from Media Processing & Delivery to Collaboration, and HPC."

Super Micro Computer, Inc. (Nasdaq: SMCI), a global leader in enterprise computing, storage, networking, and green computing technology, supports two new Intel-based accelerators for demanding cloud gaming, media delivery, AI and ML workloads, enabling customers to deploy the latest acceleration technology from Intel and Intel Habana. "Supermicro continues to work closely with Intel and Habana Labs to deliver a range of server solutions supporting Arctic Sound-M and Gaudi 2 that address the demanding needs of organizations that require highly efficient media delivery and AI training," said Charles Liang, president and CEO. "We continue to collaborate with leading technology suppliers to deliver application-optimized total system solutions for complex workloads while also increasing system performance."

Supermicro can quickly bring to market new technologies by using a Building Block Solutions approach to designing new systems. This methodology allows new GPUs and acceleration technology to be easily placed into existing designs or, when necessary, quickly adapt an existing design when needed for higher-performing components. "Supermicro helps deliver advanced AI and media processing with systems that leverage our latest Gaudi 2 and Arctic Sound-M accelerators," stated Sandra Rivera, executive vice president and general manager of the Datacenter and AI Group at Intel. "Supermicro's Gaudi AI Training Server will accelerate deep learning training in some of the fastest growing workloads in the datacenter."

At Intel's 2022 investor meeting, the company has presented a technology roadmap update to give its clients an insight into what is to come. Today, team blue announced one of the first discrete data-centric graphics cards in the lineup, codenamed Arctic Sound-M GPU. Based on the DG2 Xe-HPG variation of Intel Xe GPUs, Arctic Sound-M is the company's first design to enter the data center space. The DG2 GPU features 512 Execution Units (EUs), which get passive cooling from the single-slot design of Arctic Sound's heatsink, envisioned for data center enclosures with external airflow.

One of the most significant selling points that Intel advertises is support for hardware-based AV1 encoding standard. This feature allows the card to achieve a 30% greater bandwidth, and it is the main differentiator between consumer-oriented Arc Alchemist GPUs and itself. The card is powered by PCIe power and an 8-pin EPS power connector. Arctic Sound-M is already sampling to select customers and it will become available in the middle of 2022.Below is Intel's teaser video.