Nov. 15, 2024
Compal Unveils SG720-2A High-Performance 7U Server at SC24 with AMD technology, Ushering in a New Era of AI Computing

Compal Electronics (hereinafter referred to as Compal; Stock Ticker: 2324.TW) will showcase its cutting-edge SG720-2A AI server at the upcoming SC24 (Supercomputing 2024) conference in Atlanta, USA. Powered by AMD Instinct™ MI325X Accelerators, the SG720-2A is specifically engineered for artificial intelligence (AI) and high-performance computing (HPC) applications, setting a new standard in computational power. Designed to accommodate both EIA (19") and ORv3 (21") standards, this versatile 7U server offers unparalleled flexibility to meet the diverse requirements of modern data centers and compute-intensive workloads.

The SG720-2A combines the power of AMD EPYC™ 9005 Series processors with AMD Instinct™ MI325X GPUs UBB, delivering robust performance while optimizing space efficiency. Ideal for AI training, large-scale data analytics, and HPC workloads, the SG720-2A offers enterprises an ideal computing platform to process vast amounts of data quickly and efficiently.

The SG720-2A sets a new benchmark for high-performance computing, designed to handle the most demanding AI and HPC workloads. And it delivers advanced computing power and flexibility, enabling customers to scale operations and drive innovation.

The SG720-2A sets a new benchmark for high-performance computing, designed to handle the most demanding AI and HPC workloads. This server delivers advanced computing power and flexibility, enabling customers to scale operations and drive innovation.

Built on AMD CDNA™ 3 architecture supported by AMD ROCm™, the AMD Instinct™ MI325X Accelerators enhance the SG720-2A’s ability to accelerate key AI frameworks such as TensorFlow and PyTorch, while delivering unmatched performance for a variety of AI and machine learning tasks. The integration of these advanced components makes the SG720-2A an ideal solution for enterprises seeking to process complex AI workloads or analyze large-scale datasets in real time.

Despite its powerful computing capabilities, the SG720-2A is designed for maximum space efficiency, fitting seamlessly into a 1-meter standard rack. This space-conscious design allows data centers to maximize computational density while maintaining optimal performance, making it an ideal choice for businesses looking to scale their operations within limited physical space.

Additionally, the SG720-2A features a modular design that simplifies maintenance and reduces operational costs. Key components, including storage, UBB modules, and MB modules, can be easily swapped out without removing the entire server from the rack. This modularity streamlines system upgrades and fault-tolerant maintenance, allowing data center operators to perform these tasks without interrupting business operations. Furthermore, the network cards can be deployed without downtime, ensuring continuous operation and minimal disruption to mission-critical workloads.

The SG720-2A, powered by AMD EPYC™ 9005 Series processors, offers up to 192 “Zen 5” cores and 384 threads with clock speeds of up to 5 GHz, making it an ideal solution for data centers and cloud infrastructures. Paired with the AMD Instinct™ MI325X Accelerators, the server is optimized to handle high-performance AI computing and data-intensive applications, delivering superior computational efficiency.

Compal’s SG720-2A server will be on display at SC24, providing attendees with an opportunity to explore the latest advancements in AI computing hardware. Whether for large-scale AI model training, deep learning, or big data analytics, the SG720-2A provides businesses with the performance and scalability they need to drive the next wave of AI innovation. Looking ahead, Compal will continue its collaboration with AMD to develop next-generation GPU-powered server solutions that will continue to push the boundaries of AI and high-performance computing.

 

Media Contact  Jack Wang  Vice President and Spokesperson  02-8797-8588  investor@compal.com