ASUS has unveiled a new line of servers equipped with the AMD EPY 9005-series processors, designed to revolutionize performance and efficiency in AI-driven data center tasks.
This impressive range includes the ASUS ESC A8A-E12U, which supports AMD Instinct MI325X accelerators, and the ASUS ESC8000A-E13P, a GPU powerhouse that accommodates up to eight GPUs, making it ideal for large-scale AI model training. The ASUS RS520QA-E13, tailored for electronic design automation (EDA) and cloud computing, is another standout in the series.
For more versatile tasks, ASUS offers the RS720A-E13, RS700A-E13, RS521A, and RS501A servers, all engineered to handle demanding workloads across various applications.
At the heart of this lineup is the AMD EPYC 9005-series processor, formerly codenamed Turin, built on the Zen 5 architecture. With up to 192 cores, 384 threads, and frequencies reaching 5 GHz, these processors deliver exceptional performance for AI data centers. The EPYC 9005-series also boasts industry-leading vCPU density and enhanced compute and AI capabilities, making them ideal for AI inference tasks, whether traditional or mixed.
These servers are based on the x86 architecture, enabling seamless integration into existing x86 infrastructures for efficient migration and operational continuity in data centers.
John Morris, AMD’s corporate vice president of the Enterprise and HPC Business Group, highlighted the impact of the 5th Gen AMD EPYC processors, noting their ability to meet the rigorous demands of AI-enabled data centers. “Our high-performance compute technologies have enabled us to create a diversified enterprise portfolio that addresses the evolving needs of modern data centers,” Morris said.
Table of Contents
AI-Centric Servers: ESC A8A-E12U and ESC8000A-E13P
The ASUS ESC A8A-E12U and ESC8000A-E13P are designed specifically for AI workloads, offering support for up to eight GPUs and featuring a direct GPU-to-GPU interconnect that delivers up to 6 TB/s bandwidth.
The ESC8000A-E13P is particularly notable for its compliance with NVIDIA’s MGX architecture, which allows users to accelerate AI model training, enable real-time collaboration, and scale up infrastructure for complex projects.
Scalable Multi-Node Server: RS520QA-E13
The RS520QA-E13 is a robust server aimed at EDA and cloud computing applications. With support for advanced memory configurations and up to 400 W processors, it delivers both power and flexibility. Its ergonomic design allows for easy maintenance, while advanced thermal management ensures peak performance even in demanding environments.
General-Purpose Servers: RS700A-E13, RS720A-E13, RS521A-E12, RS501A-E12
ASUS’s general-purpose server lineup includes the RS720A-E13 and RS700A-E13, designed for AI workloads and HPC, with up to 24 DIMMs and support for powerful GPUs. The RS521A-E12 and RS501A-E12 are optimized for energy efficiency and cost-effectiveness, providing scalable solutions for data centers handling hybrid cloud environments and routine business operations.
Tailored for Diverse AI Workloads
ASUS offers a comprehensive suite of server solutions powered by AMD’s 5th Generation EPYC processors, tailored to a wide variety of AI applications, from financial computing and machine learning to computational storage and data analytics.
Availability and Pricing
ASUS’s new AMD EPYC 9005-powered servers are available globally. For more details, customers are encouraged to visit the official ASUS product page or contact their local ASUS representative.