Avalue HPS-ERSU4A/HPS-ERSUTA high-performance computing platform
Avalue Technology Inc. has introduced two new products: the HPS-ERSU4A 4U rackmount server and the HPS-ERSUTA tower server, both tailored for AI inference.
These launches come at a time when the deepening integration of digitalisation and AI across various industries boosts the demand for advanced industry-specific simulations and prototyping.
The new Avalue high-performance computing systems, the HPS-ERSU4A rackmount server and the HPS-ERSUTA tower server, are essential for addressing complex computational challenges in real-time. These systems enhance computing power and support multi-parallel models, delivering high reliability and accuracy across multiple applications, including financial market risk management, IoV vehicle telematics, and big data analytics for medical research. They are also pivotal in reducing time to market and development costs through digital twin virtual prototyping.
Constructed with Avalue HPM-ERSUA server motherboards, these servers are designed to meet the demands of the AI computing market. They are compatible with both 4th and 5th Generation Intel Xeon Scalable processors (codename Emerald Rapids), with the latter offering enhanced performance for AI inference and high-performance computing. The adoption of the 5th Generation processors also contributes to reducing environmental impact and operational costs by lowering power consumption during intensive data processing.
Additionally, the HPS-ERSU4A and HPS-ERSUTA systems feature low latency and high throughput capabilities, thanks to upgraded support for DDR5 memory DIMMs and PCIe 5.0 interfaces. These attributes make the servers particularly suitable for Cloud service providers who require fast data transmission and sophisticated data analysis.
These servers balance performance with reliability, energy efficiency, and safety. They support ODM/OEM customisation, providing options like the Intel Deep Learning Boost (Intel DL Boost) technology, which includes features such as Intel AVX-512 VNNI, AVX-512 BF16, and AMX to accelerate AI workloads. Available as either an L6 barebone system or an L10 fully assembled system, the HPS-ERSU4A and HPS-ERSUTA servers offer flexibility in configuration and rapid deployment, ultimately lowering the total cost of ownership.