Get updates delivered to you daily. Free and customizable.
Tom's Hardware
TSMC to go 3D with wafer-sized processors — CoW-SoW technology allows 3D stacking for the world's largest chips
By Anton Shilov,
2024-04-26
TSMC is taking the wafer-scale fabrication battle into the third dimension with a new technology. At its North American Technology Symposium, it introduced its next-generation system-on-wafer platform—CoW-SoW—that will enable 3D integration with wafer-scale designs. This builds on the InFO_SoW system-on-wafer integration technology technology that TSMC introduced in 2020 that allows it to build wafer-scale logic processors. So far, only Tesla has adopted this tech for its Dojo supercomputer, which TSMC says is now in production
(Image credit: TSMC)
In its forthcoming CoW-SoW platform, TSMC is set to merge two of its packaging methods — InFO_SoW and System on Integrated Chips (SoIC) — in its system-on-wafer platform. By using the Chip-on-Wafer (CoW) technology, this method will enable the stacking of memory or logic directly atop a system-on-wafer. The new CoW_SoW technology is expected to be ready for large-scale production by 2027, though it remains to be seen when actual products arrive on the market.
"So, in the future, using wafer-level integrations [will allow] our customers to integrate even more logic and memory together," said Kevin Zhang, Vice President of Business Development at TSMC. "SoW is no longer a fiction; this is something we already work with our customers [on] to produce some of the products already in place. We think by leveraging our advanced wafer-level integration technology, we can provide our customer a very important path that allows them to continue to grow their capability to bring in more computation, more energy efficient computation, to their AI cluster or [supercomputer]."
Currently, TSMC's CoW-SoW focuses on integrating wafer-scale processors with HBM4 memory. These next-generation memory stacks will feature a 2048-bit interface, which will make it feasible to integrate HBM4 directly on top of logic chips. Meanwhile, it may also make sense to stack additional logic on wafer-scale processors to optimize costs.
(Image credit: TSMC)
Wafer-scale processors in general (i.e., Cerebras's WSE), and InFO_SoW-based processors in particular, offer significant performance and efficiency benefits, including high-bandwidth and low-latency core-to-core communications, low power delivery network impedance, and high energy efficiency. As an added bonus, such processors also have additional redundancy in the form of 'extra' cores.
However, the InFO_SoW technology has certain limitations. For instance, wafer-scale processors made using this method rely entirely on on-chip memory, which may fall short for future AI needs (but is good for now). CoW-SoW will solve this, as it will allow putting HBM4 on such wafers. In addition, InFO_SoW wafers are processed using a single node, and this node does not support 3D stacking, which will be supported by CoW-SoW products.
Get updates delivered to you daily. Free and customizable.
Welcome to NewsBreak, an open platform where diverse perspectives converge. Most of our content comes from established publications and journalists, as well as from our extensive network of tens of thousands of creators who contribute to our platform. We empower individuals to share insightful viewpoints through short posts and comments. It’s essential to note our commitment to transparency: our Terms of Use acknowledge that our services may not always be error-free, and our Community Standards emphasize our discretion in enforcing policies. We strive to foster a dynamic environment for free expression and robust discourse through safety guardrails of human and AI moderation. Join us in shaping the news narrative together.
Comments / 0