China's LineShine supercomputer claims TOP500 crown with domestically designed chips

China's LineShine supercomputer dethrones US' El Capitan, secures first place in Top 500 list -- first machine in the rankings to sustain more than 2 ExaFLOPS of double-precision performance using only CPUs

China's LineShine supercomputer has dethroned El Capitan as the world's number one supercomputer, going straight to the top of the charts after the National Supercomputer Center in Shenzhen (NSCS) submitted its results. LineShine hit 2.198 FP64 ExaFLOPS in the Linpack benchmark and became the industry's first machine in the Top 500 list to sustain more than 2 ExaFLOPS of double-precision performance using only CPUs. The system is deployed at the National Supercomputing Centre in Shenzhen and was built by the Shenzhen Cloud Computing Center using semi-custom 304-core LX2 processors based on the Armv9 instruction set architecture and running at 1.55 GHz. The machine employs 13.79 million cores in total, uses proprietary LingQi interconnect, and consumes 42.2 MW of power. From a performance-per-watt point of view, the LineShine machine delivers 52.07 GFLOPS/W, which is below El Capitan's 60.94 GFLOPS/W. However, LineShine by far outperforms Fugaku -- another CPU-only supercomputer that used to be the No.1 HPC system several years ago -- that can only deliver 14.78 - 16.84 GFLOPS/W depending on whether its efficiency is optimized or not. LineShine also moved to the top of the HPCG ranking with 22.00 HPCG-PFLOPS. However, the supercomputer achieved 7.92 mixed-precision EFLOPS in HPL-MxP, which puts it behind El Capitan, Frontier, and Aurora. This limits LineShine's usability for AI training and inference, but this can be justified with its exceptional performance for traditional supercomputer tasks. Each LX2 CPU relies on two compute chiplets and has a total of 304 CPU cores organized into eight CPU clusters containing 38 cores each. Every core includes Arm SVE (Scalable Vector Extension) and SME (Scalable Matrix Extension) units that accelerate vector and matrix operations used in AI training and scientific computing that support FP64, FP32, BF16, FP16, and INT8 data formats. The chip features a rather unusual memory architecture that pairs 32 GB of on-package HBM, offering up to 4 TB/s of bandwidth with as much as 256 GB of external DDR5 memory to maximize both bandwidth and capacity. Despite this, the processor only gains 3.6X performance when moving from FP64 to mixed-precision data, which is lower compared to systems that integrate low-precision accelerators, such as AMD's Instinct MI300A or Intel's Ponte Vecchio. While an Armv9 CPU with SVE/SME can accelerate FP16/BF16/INT8 workloads, its mixed-precision uplift remains limited compared to systems with accelerators due to many reasons, including memory bandwidth, software maturity, and interconnect efficiency. That said, it may be too early to make final conclusions about the LX2 and its usability for mixed-precision workloads. In any case, the very fact that a Chinese supercomputer has achieved extraordinary FP64 performance is remarkable. Furthermore, the fact that NSCS has actually submitted results to Top 500 indicates that the organization is confident that the LineShine supercomputer relies exclusively on domestic technologies and the U.S. government cannot affect the production of these technologies. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

China just built the world's most powerful supercomputer - using Huawei chips and no GPUs

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. Chinese Syndrome: Chinese institutions have been largely absent from the TOP500 HPC rankings since 2023, when worsening US-China relations led Beijing to stop submitting its most capable systems to the list. Now, the Asian country is hitting back with a vengeance thanks to a brand-new supercomputer system ranked as the most powerful in the world. The TOP500 project has unveiled the 67th edition of its biannual ranking of the world's most powerful high-performance computing (HPC) systems. Announced at the ISC 2026 conference in Hamburg, Germany, the new list marks the surprise debut of LineShine, a previously unannounced Chinese machine that enters straight at No. 1, becoming the officially acknowledged most powerful supercomputer on the planet. LineShine posted a 2.198 Exaflop/s result on the High Performance Linpack (HPL) benchmark, which the TOP500 organization says represents roughly 80% of the system's theoretical peak performance of 2.736 Exaflop/s. It is the first HPC system in history to exceed two exaflops of sustained double-precision performance - and it achieved this milestone using an all-CPU design, with no GPU accelerators anywhere in the stack. TOP500 June 2026 - Top 5 Supercomputers The system was installed at the National Supercomputing Centre in Shenzhen (NSCS) and built around the custom "LingKun" platform, using "LX2" processors thought to be designed by Huawei. The LX2 is based on the Armv9 architecture, with each chip integrating two compute dies (304 cores in total) plus eight on-package HBM stacks delivering 32 GB of high-bandwidth memory. The platform's 13.79 million computing cores are linked by the proprietary LingQi interconnect and run China's Kylin OS, a Linux-based operating system. To put it in context, LineShine is a "made-in-China" HPC project: its processors, networking, and storage are all domestically developed, with the notable exception of relying on the Armv9 instruction set architecture, which was designed in the UK by Arm. The most outstanding result is certainly reaching first place in the TOP500 rankings, scoring an unprecedented win against the US and other major HPC superpowers. US-made El Capitan was demoted to second place (1.809 Exaflop/s), though three of the five most powerful TOP500 systems - El Capitan, Frontier, and Aurora - are still based in the US. The fifth most powerful HPC machine is Germany's JUPITER Booster, which achieved exactly 1.000 Exaflop/s on the HPL benchmark. The TOP500 team notes several remarkable takeaways from this edition. LineShine employs an all-CPU design, while other super-systems use different kinds of CPU architectures, GPUs, APUs, or even custom-made accelerators. The ultimate take is that there is no single "best" technology for performing extreme computing workloads, at least when measured by the double-precision demands of the HPL benchmark. According to the NSCS team, LineShine is the culmination of years of investment in developing domestic computing solutions independent of foreign supply chains. The system is currently being used for complex engineering simulations, scientific research, AI model development, and large-scale language model training.

World's fastest supercomputer is from China, surpassing US and Germany

One of the most interesting use cases for the supercomputer is that it is key to developing models used in artificial intelligence (AI). A supercomputer in China now outranks its US counterparts as the world's most powerful, which is the first time since 2017 that a Chinese computer has topped a list sometimes viewed as a measure of a nation's technological prowess. The LineShine computer in Shenzhen, China, displaced top-ranked US computer El Capitan in the latest version of the TOP500 ranking announced Tuesday. It was the Chinese computer's debut on the list. However, there were four European supercomputers in the top 10. Dropping to fifth place is the Jupiter supercomputer in Germany. Supercomputers are especially useful for researchers who want to collect and analyse data, as they can perform complex calculations much quicker, allowing them to address some of the world's most intricate problems, such as drug discovery, climate and weather forecasts, or modelling black holes. One of the most interesting use cases for the supercomputer is that it is key to developing models used in artificial intelligence (AI). What makes China's offering so special? Scientists behind the TOP500 project said the LineShine computer at China's National Supercomputing Center achieved 2.198 exaflops, meaning it can perform more than 2 quintillion calculations per second. El Capitan, at the US government's Lawrence Livermore National Laboratory in California, now ranks second, ahead of two other US supercomputers at national laboratories in Tennessee and Illinois. The five are the only publicly verified exascale computers in the world. LineShine differs from other high-performance computers by running entirely on conventional computer chips, or CPUs, instead of the graphics processors, or GPUs, commonly used for artificial intelligence. It relies on approximately 42.2 megawatts of electricity to operate, according to TOP500. Meanwhile, China's Premier Li Qiang on Wednesday defended the country's technological advancements as an opportunity for the world rather than a threat. Li also said the country's heavy state subsidies were not the main reason for the rapid rise of its high-tech industries, at a time when Western officials have complained that China's state support for industries from artificial intelligence to electric vehicles has provided an unfair competitive edge. China's No. 2 leader made the remarks in his speech at the opening plenary of the World Economic Forum's Annual Meeting of the New Champions, known as the "Summer Davos," held this week in the northeastern Chinese coastal city of Dalian. He acknowledged there have been growing global concerns about China's technological innovations, with some pointing to the term "China Shock 2.0," as they see the nation's high-tech boom as a threat to many advanced economies. Instead, that should be seen as "China Opportunity 2.0," he said. "From the global development perspective, 'China Opportunity 2.0' means there'll be broader access to advanced technologies and more widely shared benefits," Li said. Where does Europe stand? The five are the only publicly verified exascale computers in the world, which, though it sounds impressive, is actually just a measure of speed and how many operations per second the supercomputer can perform. Other countries with machines in the top 10 list include Italy, Switzerland, and Japan. In the top 20, Spain, Finland, the Netherlands and the United Kingdom made the list. Last year, the European Union unveiled a €20bn plan to build sites with supercomputers to develop the next generation of AI models with so-called AI gigafactories, which would link collaboration across supercomputing centres, universities and businesses.

China Beats US With World's Fastest Supercomputer, but Race Not Geared for AI Work

SAN FRANCISCO, June 23 (Reuters) - China has overtaken the U.S. to win ⁠the ⁠top spot on a list of the world's ⁠fastest supercomputers, but the results may say more about Beijing's desire to show self-sufficiency in computing systems than its standing in the global AI race, experts said. The LineShine system at the National Supercomputing Centre in Shenzhen, China, uses domestically designed chips and won the top spot on the TOP500, a biannual global ranking of supercomputers, with the country's first listing in three years. The ranking comes as the U.S. and China are increasingly competing in advanced computing, with ⁠U.S. President Donald ⁠Trump on Monday signing an executive order that aims to put the U.S. ahead of China in the emerging field of quantum computing. In the June 2026 edition of TOP500, LineShine beat out the previous titleholder, El Capitan, a supercomputer housed at Lawrence Livermore National Laboratory that the U.S. government uses to develop and maintain its nuclear weapons stockpile. But technology and policy experts interviewed by Reuters said the results do not mean that China has the world's fastest computer for AI work because of changes in the computing industry in recent years and the methods ⁠used ⁠to compile the list. LineShine ranked fourth ⁠on a benchmark test designed to simulate computing work that is more similar to AI. BENCHMARK TESTS For decades, supercomputers strung together many separate machines to work on complex scientific problems such as simulating how atoms ⁠interact with one another and were mostly the domain of national labs and universities. To be ranked on the TOP500 list, supercomputer operators must run a set of benchmark tests that aims to mimic such work. But in more recent years, cloud computing companies such as Microsoft, Amazon.com and Alphabet's Google built out massive supercomputers of their own but geared them for AI work instead. Most of those companies do not opt to compete for a spot on the TOP500 ⁠list. A study last year by AI policy researchers Konstantin Pilz, James Sanders, Robi Rahman and Lennart Heim found ⁠that SpaceX-owned xAI's Colossus system was already likely more powerful than the U.S. government's El Capitan. "If the hyperscalers submitted their systems, this 'world's fastest' would not crack the top five," said Jimmy Goodrich, a senior fellow at the University of California's Institute for Global Conflict and Cooperation. CHIP DESIGN EFFORTS The Chinese victory on the list more likely shows that China wanted recognition for its chip design efforts, which is a change from recent years, experts said. China first took the top spot on the TOP500 in 2010 and traded titles back and forth with the U.S. and Japan until 2023, when China stopped submitting its systems after years of chip- and computing-related export controls from Trump's first administration and later under President Joe Biden. "I'm not surprised it's the number one system. What I'm surprised by is ⁠that they submitted it and want recognition for it," said Addison Snell, CEO of Intersect360 Research, a firm that focuses on supercomputers. The LineShine system does not contain any advanced AI chips, according to details presented with the results, likely because the tools to make those chips are still subject to U.S. export controls. "China is hoping to convince the world export controls are useless by hoping we ignore the details," Goodrich said. The National Supercomputing Centre did not respond immediately to a request for comment. (Reporting by Stephen Nellis in San Francisco; Editing by Peter Henderson and Jamie Freed)

In supercomputer marathon, China tops US; but AI-high? No

China's LineShine supercomputer has topped the global TOP500 list, utilizing domestically produced chips. However, experts suggest this achievement highlights China's drive for self-sufficiency rather than a definitive lead in AI computing. While LineShine secured the top spot, it ranked lower on AI-specific benchmarks, indicating the evolving landscape of high-performance computing beyond traditional scientific simulations. San Francisco: China has overtaken the US to win the top spot on a list of the world's fastest supercomputers, but the results may say more about Beijing's desire to show self-sufficiency in computing systems than its standing in the global AI race, experts said. The LineShine system at the National Supercomputing Centre in Shenzhen, China, uses domestically designed chips and won the top spot on the TOP500, a biannual global ranking of supercomputers, with the country's first listing in three years. The ranking comes as the US and China are increasingly competing in advanced computing, with US President Donald Trump on Monday signing an executive order that aims to put the US ahead of China in the emerging field of quantum computing. In the June 2026 edition of TOP500, LineShine beat out the previous titleholder, El Capitan, a supercomputer housed at Lawrence Livermore National Laboratory that the US government uses to develop and maintain its nuclear weapons stockpile. But technology and policy experts said the results do not mean that China has the world's fastest computer for AI work because of changes in the computing industry in recent years and the methods used to compile the list. LineShine ranked fourth on a benchmark test designed to simulate computing work that is more similar to AI. BENCHMARK TESTS For decades, supercomputers strung together many separate machines to work on complex scientific problems such as simulating how atoms interact with one another and were mostly the domain of national labs and universities. To be ranked on the TOP500 list, supercomputer operators must run a set of benchmark tests that aims to mimic such work. But in more recent years, cloud computing companies such as Microsoft, Amazon. com and Alphabet's Google built out massive supercomputers of their own but geared them for AI work instead. Most of those companies do not opt to compete for a spot on the TOP500 list. A study last year by AI policy researchers Konstantin Pilz, James Sanders, Robi Rahman and Lennart Heim found that SpaceX-owned xAI's Colossus system was already likely more powerful than the U.S. government's El Capitan.

China beats U.S. with world's fastest supercomputer, but race not geared for AI work

SAN FRANCISCO - China has overtaken the U.S. to win the top spot on a list of the world's fastest supercomputers, but the results may say more about Beijing's desire to show self-sufficiency in computing systems than its standing in the global AI race, experts said. The LineShine system at the National Supercomputing Center in Shenzhen, China, uses domestically designed chips and won the top spot on the TOP500, a biannual global ranking of supercomputers, with the country's first listing in three years. The ranking comes as the U.S. and China are increasingly competing in advanced computing, with U.S. President Donald Trump on Monday signing an executive order that aims to put the U.S. ahead of China in the emerging field of quantum ⁠computing. In the June 2026 edition of TOP500, LineShine beat out the previous titleholder, El Capitan, a supercomputer housed at Lawrence Livermore National Laboratory ⁠that the U.S. government uses to develop and maintain its nuclear weapons stockpile. But technology and policy experts said the results do not mean that China has the world's fastest computer for AI work because of changes in the computing industry in ⁠recent years and the methods used to compile the list. LineShine ranked fourth on a benchmark test designed to simulate computing work that is more similar to AI. For decades, ⁠supercomputers strung together many separate machines to work on complex scientific problems such as simulating how atoms interact with one another and were mostly the domain of national labs and universities. To be ranked on the TOP500 list, supercomputer operators must run a set of benchmark tests that aims to mimic such work. But in more recent years, cloud computing companies such as Microsoft, Amazon.com and Alphabet's Google built out massive supercomputers of their own but geared them for AI work instead. Most of those companies do not opt to compete for a spot on the TOP500 list. A study last year by AI policy researchers Konstantin Pilz, James Sanders, Robi Rahman and Lennart Heim found that SpaceX-owned xAI's Colossus system was already likely more powerful than the U.S. government's El Capitan. "If the hyperscalers submitted their systems, this 'world's fastest' would not crack the top five," said Jimmy Goodrich, a senior fellow at the University of California's Institute for Global Conflict and Cooperation. The Chinese victory on the list more likely shows that China wanted recognition for its chip design efforts, which is a change from recent years, experts said. China first took the top spot on the TOP500 in 2010 and traded titles back and forth ⁠with the U.S. and Japan until 2023, when China stopped submitting its systems after years of chip- and computing-related export controls from Trump's...

China Builds Fastest Supercomputer, Beats a US Machine for the First Time Since 2017

China has claimed that it now has the world's fastest supercomputer, taking over that coveted crown from the United States at a time when Washington and Beijing are amidst a fierce technological competition around artificial intelligence and chipmaking. Both claim that they've a right to keep the other out for the sake of their own national security. Per a report published by The New York Times, the LineShine computing system in Shenzhen province of China has been declared the fastest computing system by researchers who used a set of standard tests designed for supercomputers. Dr. Jack Dongarra, a professor of computer science and electrical engineering at the University of Tennessee, who led a team to inspect the supercomputer says the underlying design achieved by the scientists in China points to a better way to blend AI with traditional scientific tasks. Incidentally, the doctor is the one to organised the Top-500 list of supercomputers. According to the report, LineShine not only showcased raw speed, but also stood out due to its use of standard microprocessors. Coming at a time when every single chipmaker in the world is harping on special chips (also called GPUs) that is considered the lifeline of AI, this solution from the Chinese company might just cause a few flutters. According to Dr. Dongarra, the test results seen at the Shenzhen Cloud Computing Centre for LineShine was over 20% faster than those of El Capital, a supercomputing system located at the Lawrence Livermore National Laboratory in California. This computing system topped a twice-yearly ranking of supercomputer performance since November 2024. Significantly, this is the first time since 2017 that China has managed to usurp the top position in terms of supercomputers from the US and before them Japan. Dr. Dongarra found the system impressive and felt that the Chinese had upped the United States by developing a system that is not reliance in GPUs. The Top500 website revealed the results of the latest test for the world's most powerful supercomputers during the ISC 2026 conference at Hamburg, Germany. "LineShine, a previously unlisted system installed in China, debuts at No. 1, displacing El Capitan as the world's most powerful supercomputer as measured by the High Performance Linpack (HPL) benchmark. The new list also reflects continued depth in U.S. and European exascale capability, a new entrant in Italy's HPC fleet, and unchanged leadership atop the Green500 energy-efficiency ranking," the website said. Of course, it is anybody's guess as to how things would pan out between China and the US whose battle for tech supremacy has seen President Donald Trump impose restrictions of all kinds on exporting local knowhow. However, the fact remains that China has trodden its own path despite these hurdles that most likely helped US tech giants like Anthropic, OpenAI and Google develop frontier AI models while the likes of Nvidia and a now government-owned Intel building AI chips. The fact that Microsoft is seeking out DeepSeek's AI model as an option for its Copilot-led enterprise AI business is itself a win for Beijing. All that this new race for the most powerful supercomputer between the two rivals does would be in the form of red signals flashing at the White House, which has already imposed tariffs and placed limits on AI chip exports, including those made by Nvidia. Now that China has used CPUs to power a supercomputer could blow hot on the Trump administration as well as the tech czars who are following the Pied Piper of Washington DC. Would America now attempt to restrict export of CPUs to the Chinese market? Well, time will tell if this is so and how Beijing would respond to such a threat. In fact, Chinese CPU industry has set about transforming itself for tech self-sufficiency backed by a massive state funding to the tune of $47.5 billion. Top fabless designers and foundries are said to be narrowing the performance gap with the global giants such as AMD and Intel. Key chip designers like Loongson, Huawei and Zhaoxin have used proprietary architecture for a series of use cases, ranging from desktop and servers to mobile and government uses. Similarly in the field of manufacturing and fabrication, the Semiconductor Manufacturing International Corporation (SMIC) has functioned as their premier foundry despite not having access to the advanced extreme ultraviolet lithography equipment. Reports indicate that SMIC has achieved robust performances using the 7nm processes with 5nm in active development. Of course, it is another matter that the supercomputer edge today does not really count for much given that new age commercial AI systems operated by the likes of Google can prove to be faster. The use of approximations for tasks like selecting the next word or identifying images using 4-bit and 8-bit numbers makes them far more efficient. Each time China or Japan have overtaken them in the supercomputer race, the administration has pushed for more funding, the last such being in the form of the Genesis Mission started by the Trump administration last November. The aim is to exploit all the supercomputers in the US national labs and private companies to supercharge AI. What China seems to have done though is to ensure that LineShine does not separate the traditional jobs of CPUs and GPUs, instead it creates tasks with specialised circuitry for the latter and accelerates matrix and vector calculations. That ability is embedded in chips that have a total of nearly 14 million computing cores, or tiny electronic brains, installed in 90 hardware cabinets, says the article in NYT. So, in the end does it mean this may end up as a Pyrrhic victory for China? The scientists at Shenzhen do not think so. For, they have sought global recognition for their supercomputer through submissions for the Gordon Bell Prize, which is considered a benchmark for sophisticated computing.