Extreme Measures Needed to Scale Chips
The July 2024 IEEE Spectrum issue discusses scaling compute power for AI, exploring solutions like EUV lithography, linear accelerators, and chip stacking. Industry innovates to overcome challenges and inspire talent.
Read original article
In the July 2024 issue of IEEE Spectrum, the focus is on scaling compute power to meet the demands of Artificial Intelligence (AI). As Dennard scaling, the traditional method of increasing transistor density, began losing effectiveness in the mid-2000s, chipmakers turned to solutions like extreme ultraviolet lithography. Japanese researchers at KEK are exploring the use of a linear accelerator as an EUV light source to push the boundaries of chip technology. While the industry faces challenges in making smaller devices economically, innovations like brighter light sources and future transistors offer hope. Stacking chips is seen as a practical way to enhance logic and memory capabilities. TSMC is betting on stacking GPUs to meet the needs of the AI sector. The company aims to inspire young talent to pursue semiconductor engineering despite concerns about the end of Moore's Law. Overall, the industry is exploring various strategies to continue advancing semiconductor technology and meet the growing demands of AI applications.
Related
TSMC experimenting with rectangular wafers vs. round for more chips per wafer
TSMC is developing an advanced chip packaging method to address AI-driven demand for computing power. Intel and Samsung are also exploring similar approaches to boost semiconductor capabilities amid the AI boom.
Etched Is Making the Biggest Bet in AI
Etched invests in AI with Sohu, a specialized chip for transformers, surpassing traditional models like DLRMs and CNNs. Sohu optimizes transformer models like ChatGPT, aiming to excel in AI superintelligence.
Sohu: The First Transformer ASIC
Etched secures $120 million to develop Sohu, a transformer ASIC enhancing AI model performance. Sohu enables real-time voice agents, rapid text processing, and trillion-parameter models, revolutionizing AI processing with advanced features.
Sohu AI chip claimed to run models 20x faster and cheaper than Nvidia H100 GPUs
Etched startup introduces Sohu AI chip, specialized for transformer models, outperforming Nvidia's H100 GPUs in AI LLM inference. Sohu aims to revolutionize AI processing efficiency, potentially reshaping the industry.
Extreme Measures Needed to Scale Chips
Semiconductor technology faces challenges in scaling for AI demands. Innovations like EUV lithography and chip stacking are key. Japanese researchers explore linear accelerators for EUV light. Attracting young talent is crucial.
3 commentsPerhaps he should try paying them accordingly.
Hardware companies don’t pay well, see salaries at places like Broadcom, Qualcomm, Intel and IBM, at least compared to top software companies (and even compared to jobs such as in accounting or project management). That’s why.
It's interesting how chip designs are mimicking terminator brain chip design and real estate (limits of horizontal expansion of designs force them to go up).
Related
TSMC experimenting with rectangular wafers vs. round for more chips per wafer
TSMC is developing an advanced chip packaging method to address AI-driven demand for computing power. Intel and Samsung are also exploring similar approaches to boost semiconductor capabilities amid the AI boom.
Etched Is Making the Biggest Bet in AI
Etched invests in AI with Sohu, a specialized chip for transformers, surpassing traditional models like DLRMs and CNNs. Sohu optimizes transformer models like ChatGPT, aiming to excel in AI superintelligence.
Sohu: The First Transformer ASIC
Etched secures $120 million to develop Sohu, a transformer ASIC enhancing AI model performance. Sohu enables real-time voice agents, rapid text processing, and trillion-parameter models, revolutionizing AI processing with advanced features.
Sohu AI chip claimed to run models 20x faster and cheaper than Nvidia H100 GPUs
Etched startup introduces Sohu AI chip, specialized for transformer models, outperforming Nvidia's H100 GPUs in AI LLM inference. Sohu aims to revolutionize AI processing efficiency, potentially reshaping the industry.
Extreme Measures Needed to Scale Chips
Semiconductor technology faces challenges in scaling for AI demands. Innovations like EUV lithography and chip stacking are key. Japanese researchers explore linear accelerators for EUV light. Attracting young talent is crucial.