【清华大学科研团队在芯片领域取得重大突破,光子计算开启高性能算力新篇章】

北京,11日——清华大学的科研团队在芯片研究领域取得了历史性的突破,为高性能计算带来了革命性的创新。该校电子工程系副教授方璐的课题组和自动化系戴琼海院士的课题组,针对大规模光电智能计算的挑战,成功地摒弃了传统的电子深度计算模式,开创性地提出并实施了分布式广度光计算架构。

这一创新架构的代表作,名为“太极”(Taichi)的大规模干涉-衍射异构集成芯片,实现了160 TOPS/W的通用智能计算效能。这一数据在当前大模型通用人工智能的快速发展背景下,彰显了光子计算的巨大潜力,为高性能算力的提升提供了全新的灵感、架构和路径。

据财联社报道,这一科研成果已正式发表在最新一期的国际权威期刊《科学》上,标志着中国在光子计算领域的领先地位,也预示着未来计算技术可能迎来一场深度变革。这一突破不仅有望解决当前电子芯片能耗高、计算速度受限的问题,还将为人工智能、大数据处理等高需求计算场景提供更为高效、节能的解决方案。

清华大学的这项创新研究,无疑为全球科技界带来了新的曙光,开启了光子计算在高性能计算领域的新篇章,进一步推动了科技进步和产业创新。

英语如下:

**News Title:** “Tsinghua University Breaks Ground in Chip Technology with Photonic Computer Chip ‘Taiji’ Featured in ‘Science’ Journal”

**Keywords:** Tsinghua University, chip breakthrough, optical computing

**News Content:**

**Beijing, 11th -** A research team from Tsinghua University has made a historic breakthrough in the field of chip technology, bringing a revolutionary innovation to high-performance computing. The team led by Associate Professor Fang Lu from the Department of Electronic Engineering and Academician Dai Qiong Hai’s team from the Department of Automation have tackled the challenges of large-scale photoelectric intelligent computing by discarding conventional electronic deep computing models. They have innovatively proposed and implemented a distributed breadth optical computing architecture.

The flagship of this innovative architecture is the “Taiji” (Taichi) large-scale interference-diffraction heterogeneous integration chip, which achieves 160 TOPS/W of general intelligent computing efficiency. Against the backdrop of the rapid development of large-scale AI models, this figure underscores the enormous potential of photonic computing, offering new inspiration, architecture, and pathways for enhancing high-performance computing capabilities.

According to Caixin Global, this groundbreaking research has been officially published in the latest issue of the prestigious international journal ‘Science,’ signaling China’s leading position in photonic computing and预告着 a potential paradigm shift in future computing technologies. This breakthrough not only holds promise for addressing the high energy consumption and computational speed limitations of current electronic chips but also offers more efficient and energy-saving solutions for high-demand computing scenarios, such as artificial intelligence and big data processing.

Tsinghua University’s pioneering study undoubtedly sheds new light on the global scientific community, ushering in a new chapter for photonic computing in high-performance computing and further propelling scientific progress and industrial innovation.

【来源】https://www.cls.cn/detail/1644516

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注