ASRock Rack, a provider of server technology aimed at HPC, Cloud, and large-scale data center users, delivered its new midrange server to us for testing. The ASRock Rack 2U12L-ICX2 is an affordable barebone server designed for small to medium businesses and is well suited to System Integrators and Value-added Resellers. The company is known for

Dell Technologies used SC22 to announce additions to the PowerEdge server portfolio, HPC workloads on APEX through an as-a-Service experience, integrating hybrid quantum computing into existing classical computational ecosystems, and Dell Validated Design for HPC for Risk Assessment that analyzes large datasets.

AMD has announced the general availability of the new EPYC 9004 4th-gen CPUs. Code-named AMD Genoa, the new line of CPUs supports 12 channels of DDR5-4800 (up to 6TB memory capacity per socket), 128 lanes of PCIe Gen5, AMD Infinity Fabric/Guard technology, and up to 96 cores. This makes them ideal for critical workloads across

The Lenovo ThinkSystem ST250 V2 is a single-socket server for small businesses supporting on-premise setups. It’s ideal for office tasks, especially filesharing and surveillance because of its highly versatile storage options. Configurations can include 16 2.5-inch, eight 2.5-inch, or a mix of 2.5- and 3.5-inch drives which support simple or hot-swapping. Despite its budget positioning,

GIGABYTE Technology has announced the R182-P91, R282-P91, R282-P92 and H262-P61 high-density, ARM-based compute servers, supporting up to a massive 256 CPU cores. These new servers are designed for cloud-native applications using Ampere Altra processors that support dual socket configurations.

EK is adding to their Fluid Works Compute Series with a new rackmount liquid-cooled GPU server, the X7000-RM. This new liquid-cooled server is purpose-built for high-compute density applications including machine learning, artificial intelligence, rendering farms, and scientific compute simulations.

Edge servers facilitate real-time decision-making by providing computing resources away from data centers and the cloud. In this article, we run several edge benchmarks on a SuperMicro IoT SuperServer SYS-210SE-31A, a versatile multi-node edge server. Clearly, the engineers had edge inferencing in mind during its conception, as this box was born to inference.

Even more versatile in its second generation, the Lenovo ThinkSystem SR650 V2 is a mainstream server ideally suited for cloud storage, VDI, HPC, compute, and enterprise applications. This 2U, two-socket server is configurable with 40 storage drives or eight GPUs and 8TB of RAM.

Technology moves in cycles, and no cycle is more evident right now than the emphasis on AI at the edge. In particular, we’re finding a massive swing to edge inferencing. NVIDIA is a big part of this push, wanting to drive adoption of their GPUs outside the data center. Still, the fact is that enterprises