Bigstack Partners with NYCU to Build an IC Design Education Cloud

Source: DIGITIMES

As advanced semiconductor processes such as FinFET and 7nm have become increasingly prevalent across the semiconductor industry, talent development has evolved beyond traditional circuit design to encompass large-scale IC design implementation that meets advanced process requirements and cloud-based security standards. Enabling students to gain hands-on experience with industry-relevant EDA (Electronic Design Automation) tools and advanced process design environments while still in school has therefore become an important challenge for universities.

With support from the Ministry of Education’s Semiconductor Program, National Yang Ming Chiao Tung University (NYCU) deployed high-performance servers powered by AMD EPYC^TM processors and partnered with Bigstack to build an EDA/HPC (High Performance Computing) education cloud platform powered by CubeCOS. The private cloud environment is designed to support more than 500 students simultaneously, bringing advanced process design education into the classroom.

Beyond Advanced Process Education: The Challenge Goes Beyond Computing Power

Professor Ming-Dou Ker, Associate Dean of the Department of Electronics and Electrical Engineering at NYCU, noted that as courses increasingly incorporate advanced processes such as FinFET, the EDA tools and design workflows students use are becoming increasingly similar to those found in the semiconductor industry. This not only significantly increases computing requirements, but also places greater demands on environment management and security.

Because process models, design rules, and PDKs (Process Design Kits) contain highly sensitive information, the university needed a private cloud platform capable of providing both workload isolation and centralized management.

Dr. Lin-Hung Lai, a doctoral graduate of NYCU’s Institute of Electronics who was responsible for planning the system, pointed out that the biggest challenge was not hardware, but the manpower required to manage it. “Our biggest problem was actually not the hardware, but the manpower.” In the past, graduate students had to handle system maintenance in addition to their research work. A failure on a single server could potentially affect hundreds of students taking classes, making the reduction of operational complexity a key priority in the development of the new platform.

CubeCOS Enables a Scalable Cloud Platform for Advanced IC Design Education

With CubeCOS as the cloud platform, NYCU consolidated previously distributed computing resources into a unified resource pool and created multi-tenant, isolated environments that balance performance and security requirements according to course needs. This allows sensitive process data, including PDKs, to remain appropriately isolated across different course environments. When an individual node encounters an issue, maintenance can be performed on the affected environment without disrupting the overall teaching services, significantly improving system availability and operational efficiency.

Beyond resource management, Bigstack’s open APIs also enabled NYCU to develop its own Account Management System (AMS). The system automates processes including account applications, account provisioning, access management, and password resets.

According to Dr. Lai, API integration has transformed administrative processes that previously relied heavily on manual work into automated software workflows, allowing the limited IT resources to focus more on teaching support and technical services.

NYCU has currently deployed 10 high-performance servers, centrally managed through CubeCOS, supporting 18 IC-related courses and more than 500 students conducting design and simulation exercises concurrently. The platform also provides room for future expansion to support more complex IC design courses and AI applications.

For Dr. Lai, one of the most valuable outcomes of the project was not simply the adoption of a new platform, but the collaborative approach to solving technical challenges with a Taiwan-based technology team. “From hardware to software, we can communicate directly with the engineers to discuss problems, and many issues can be resolved in a very short time.” He believes this close collaboration has produced more than technical results—it has also helped foster an ecosystem in which education and industry can grow together.

Professor Ko summed up the core vision of the project in one sentence:

“We are not providing an experimental environment. We are building a laboratory.”

From establishing the education cloud platform to continuously expanding its capabilities for AI and advanced-process computing, NYCU aims to build more than a system that simply supports coursework. It is developing a long-term semiconductor talent development infrastructure that can support education, research, and industry collaboration, while providing a replicable model for digital transformation in higher education in Taiwan.