Quanta Cloud Technology (QCT), Supermicro, Giga Computing Technology (GCT), Mitac Computing Technology (MCT), and Ingrasys Technology, a Foxconn Industrial Internet (Fii) subsidiary, are all showcasing AI servers and other innovative products at the ongoing Computex 2023 held in Taipei, Taiwan.
Quanta Computer subsidiary QCT is exhibiting 5G, high-performance computing (HPC), AI, and cloud solutions at the show. The company said it is committed to creating customized 5G and AI infrastructure to help customers face future challenges.
Exhibition highlights from QCT include its data center and edge hardware, 5G enterprise private network equipment, and liquid cooling racks.
Mitac Holdings subsidiary MCT is showcasing HPC, cloud computing, storage, and Open Compute Project (OCP) server solutions. Its newest products include edge servers for 5G environments, which can provide telecom operators with the infrastructure to build 5G base stations.
Eric Kuo, VP of the server infrastructure business unit at MCT pointed out that the demand for new technology and big data analysis applications has increased, driving HPC and cloud computing upgrades. Continuous breakthroughs in computing power are also accelerating the development of HPC and AI applications.
MCT is showing server platforms supporting the latest Epyc 9004 series processors from Advanced Micro Devices (AMD) and HPC, cloud computing, and storage platforms powered by 4th generation Intel Xeon scalable processors from its subsidiary TYAN.
These platforms include a 2U multi-node server, a 26-disk full-flash storage server, a cost-effective 1U dual-socket cloud server, and a 2U virtualized application server that supports 32 DDR5 RDIMMs.
TYAN's Transport HX TN85-B8261 high-performance server supports up to four double-wide GPU cards and can meet the requirements to carry out HPC and deep learning applications.
TYAN's Thunder HX FT65T-B5652 and Transport HX FT65T-B8050 models are rack-convertible pedestal server platforms suited for desk-side AI workbench workloads.
In response to increasing energy costs, MCT has developed two OPC servers based on Open Rack V3 architecture: Capri2 CP2S11 and Goldstone GS1D01.
The Capri2 CP2S11 is compatible with Open Rack V2 and V3 architectures and can be used with Epyc 9004 processors, as well as double-wide GPU cards, to meet the most basic computing needs for AI applications.
The Goldstone GS1D01 supports 4th generation Intel Xeon scalable processors and is equipped with 32 DDR5 memory slots, ideal for computing platforms with large-capacity memory needs.
MCT is also offering a solution for 5G O-RAN base stations with its Whitestone WS1S01 distributed unit (DU) server, which has built-in IEEE 1588 V2 precision time protocol (PTP) hardware and firmware and provides up to eight 10G SFP+ and four 25G SFP2 network ports.
Supermicro is exhibiting a full range of IT solutions for cloud, AI, machine learning, storage, 5G, and smart edge applications.
Supermicro founder and CEO Charles Liang stated that his company will continue to work closely with Nvidia to introduce energy-efficient servers for AI and other industries.
Supermicro currently ships 4,000 racks per month and expects to ship more than 5,000 per month by the end of the year, according to Liang.
Supermicro's all-new server series based on Nvidia's Grace CPU superchip will hit the market soon. Each new server will come equipped with 144 cores with dual CPUs joined by a 900GB per second connection, good for realizing highly responsive AI applications and applications requiring ultra-low latency responses.
Gigabyte Technology subsidiary GCT is showcasing a range of enterprise and consumer products, including servers supporting the Nvidia Grace CPU superchip platform. GCT is also exhibiting servers supporting Intel, Nvidia's HGX H100 system, Ampere Computing's AmpereOne platform, and AMD's Alveo V70 AI inference accelerator cards. Also on display are liquid-cooled servers with direct liquid cooling (DLC) technology.
Article translated by Eifeh Strom
