In an era where artificial intelligence (AI) is increasingly woven into the fabric of various industries, the demand for computing power is reaching unprecedented heights. Chu Junhao, a member of the Chinese Academy of Sciences and a renowned semiconductor specialist, highlights the critical need to address this surge in computing requirements.
“As AI continues to integrate with sectors like medicine, materials science, information technology, energy, and health, we are witnessing the birth of numerous new industrial opportunities,” Chu explains. “However, these advancements hinge on our ability to provide substantial computing power.”
Chu emphasizes the strategic enhancement and expansion of computational infrastructure. “We must carefully plan the distribution of computing power, both regionally and within its internal structure,” he says. “Enhancing computing power through advancements in semiconductor hardware technology is essential.”
Innovations such as compute-in-memory (CIM) technology are pivotal in this endeavor. By enabling on-chip processing, CIM reduces the need to transmit large quantities of unprocessed data, thereby significantly decreasing storage demands. “Processing data on-chip allows us to isolate smaller, more manageable units of information,” Chu notes. “Organizing data into units facilitates subsequent processing stages and conserves computing resources.”
The deceleration of Moore’s Law—which predicted the doubling of processing power every two years—signals a turning point in technological development. As semiconductor components near atomic scales, sustaining this growth requires new materials and innovative computing models. Chu points to spintronic devices, quantum chips, and photonic computing as future game-changers. “We are exploring the scientific principles and practical applications of these technologies,” he says. “Their full implementation will open new avenues for real-world applications.”
Beyond hardware advancements, Chu underscores the importance of incorporating human-like intelligence into computing systems. He cites the example of a chess-playing robot that consumes vast amounts of energy compared to the human brain. “Embedding the scientific principles of human thought into computers is a fundamental aspect of AI’s future development,” he asserts. “It’s a gradual journey that requires thorough research and won’t yield instant success.”
As the quest to meet the increasing computing demands continues, Chu’s insights underscore the necessity of innovation across multiple fronts. From strategic infrastructure planning to breakthroughs in hardware and artificial intelligence, these efforts are crucial in navigating the challenges of our rapidly advancing digital age.
Reference(s):
Tech Talk: How to cope with the increasing need of computing power?
cgtn.com
