ASE Embraces AMD CPUs & MI300 GPUs for AI
ASE Technology is harnessing the power of AMD EPYC processors and Ryzen CPUs, resulting in a notable performance boost and considerable energy savings. The company, a leader in semiconductor assembly and test, reports a 50% performance increase and a 6.5% reduction in power consumption thanks to this strategic shift. They are also actively evaluating AMD Instinct MI300 for demanding AI workloads. This move underscores ASE’s commitment to cutting-edge technology in its data centers and client systems. Discover how these advancements lead to lower operational costs and improved capabilities, positioning ASE for continued growth. Furthermore, observe their collaboration with AMD on advanced 2.5D packaging and their exploration of high-bandwidth memory solutions. News Directory 3 is following this story closely. Discover what’s next for ASE’s AI endeavors.
AMD EPYC Processors Drive Performance Gains for ASE Technology
ASE Technology holdings,the world’s largest outsourced semiconductor assembly and test provider,is seeing significant benefits from its transition to AMD EPYC processors and Ryzen CPUs. The company is leveraging these processors across its data centers and client systems to achieve improved performance and energy efficiency.
The move to AMD EPYC and Ryzen has resulted in a 50% increase in system performance and a 6.5% reduction in power consumption, according to AMD. This translates to a 30% decrease in total cost of ownership for ASE. The company is also evaluating AMD’s Instinct MI300-series processors for AI workloads.
Jekyll Chen, director of IT infrastructure for ASE, emphasized the need for high performance, low latency, and high core counts to handle large volumes of data analysis. “We need to handle a big volume of data analysis, including leading-edge technology for AI applications and our smart factories,” Chen said. “Stability and scalability are two primary goals for us.”
ASE has collaborated with AMD on advanced 2.5D packaging since 2007, contributing to the advancement of high-bandwidth memory (HBM). While ASE provides packaging services for AMD, it remains unclear whether they package AMD’s AI GPUs, which utilize TSMC’s CoWoS technology.
AMD reports growing adoption and evaluation of its Instinct processors for on-prem AI inference. ASE’s evaluation of the Instinct