Latest news with #LawrenceBerkeleyLab
New York Times
31-07-2025
- Science
- New York Times
Mary K. Gaillard, Physicist Who Probed the Subatomic Universe, Dies at 86
Mary K. Gaillard was 16 and still known as Mary Ralph when a boy in her neighborhood asked her what she wanted to do with her life. She told him that she wanted to be a physicist. 'A singularly unfeminine profession,' he replied. Decades later, that remark would inspire the title of Dr. Gaillard's memoir, 'A Singularly Unfeminine Profession: One Woman's Journey in Physics' (2015), in which she recounted a career spanning a golden age of particle physics, when the outlines of how nature behaves at subatomic scales were just beginning to emerge. Dr. Gaillard contributed key insights to what is now known as the Standard Model — scientists' best theory about the properties and interactions of elementary particles — while overcoming discrimination as one of the few women in her field and inspiring other female physicists to do the same. Physics was 'her life,' her son Bruno said. 'She was consumed by it.' Known to many as Mary K, sans period, Dr. Gaillard, who died on May 23 at 86, was the first woman hired by the physics department at the University of California, Berkeley, and later became a senior scientist at Lawrence Berkeley National Laboratory. But much of her groundbreaking work occurred earlier, during a long stint as an unpaid visiting scientist at the European Organization for Nuclear Research, or CERN, a laboratory on the Franco-Swiss border. She was 'brilliant at doing calculations,' said John Ellis, a physicist at King's College London, who collaborated with Dr. Gaillard at CERN. 'If she calculated something, you could be sure that it was correct.' Want all of The Times? Subscribe.
Reuters
10-06-2025
- Business
- Reuters
Nvidia, HPE to build new supercomputer in Germany
SAN FRANCISCO, June 10 (Reuters) - Nvidia (NVDA.O), opens new tab and Hewlett Packard Enterprise (HPE.N), opens new tab said on Tuesday they are partnering with the Leibniz Supercomputing Centre to build a new supercomputer using Nvidia's next-generation chips. The Blue Lion supercomputer, as the project is called, will become available to scientists in early 2027, using Nvidia's "Vera Rubin" chips. The announcement, made at a supercomputing conference in Hamburg, Germany, follows Nvidia's announcement that the Lawrence Berkeley National Lab in the United States also plans to build a system using the chips next year. Separately, Nvidia also said that Jupiter, another supercomputer using its chips at German national research institute Forschungszentrum Julich, has officially become Europe's fastest system. The deals represent European institutions aiming to stay competitive against the U.S. in supercomputers used for scientific fields from biotechnology to climate research. Long before it became an artificial intelligence powerhouse, Nvidia set out to persuade scientists to use its chips to speed up complex computer problems, such as modeling climate change. Those problems required many precise calculations that could take months at a time. Nvidia is now working to persuade scientists to use artificial intelligence. Those AI systems can take the results of a few precise calculations and use them to make predictions that, while not as accurate as the fully calculated results, can still be useful while taking far less time. Nvidia on Tuesday unveiled what it calls its "Climate in a Bottle" AI model. In a press briefing, Dion Harris, head of data center product marketing at Nvidia, said scientists will be able to input a few initial conditions such as sea surface temperatures and generate a forecast for 10 to 30 years in the future and see what the weather may be like at any kilometer or so of the earth's surface. "Researchers will use combined approach of classic physics and AI to resolve turbulent atmospheric flows," Harris said. "This technique will allow them to analyze thousands and thousands more scenarios in greater detail than ever before."
Yahoo
10-06-2025
- Business
- Yahoo
Nvidia, HPE to build new supercomputer in Germany
By Stephen Nellis SAN FRANCISCO (Reuters) -Nvidia and Hewlett Packard Enterprise said on Tuesday they are partnering with the Leibniz Supercomputing Centre to build a new supercomputer using Nvidia's next-generation chips. The Blue Lion supercomputer, as the project is called, will become available to scientists in early 2027, using Nvidia's "Vera Rubin" chips. The announcement, made at a supercomputing conference in Hamburg, Germany, follows Nvidia's announcement that the Lawrence Berkeley National Lab in the United States also plans to build a system using the chips next year. Separately, Nvidia also said that Jupiter, another supercomputer using its chips at German national research institute Forschungszentrum Julich, has officially become Europe's fastest system. The deals represent European institutions aiming to stay competitive against the U.S. in supercomputers used for scientific fields from biotechnology to climate research. Long before it became an artificial intelligence powerhouse, Nvidia set out to persuade scientists to use its chips to speed up complex computer problems, such as modeling climate change. Those problems required many precise calculations that could take months at a time. Nvidia is now working to persuade scientists to use artificial intelligence. Those AI systems can take the results of a few precise calculations and use them to make predictions that, while not as accurate as the fully calculated results, can still be useful while taking far less time. Nvidia on Tuesday unveiled what it calls its "Climate in a Bottle" AI model. In a press briefing, Dion Harris, head of data center product marketing at Nvidia, said scientists will be able to input a few initial conditions such as sea surface temperatures and generate a forecast for 10 to 30 years in the future and see what the weather may be like at any kilometer or so of the earth's surface. "Researchers will use combined approach of classic physics and AI to resolve turbulent atmospheric flows," Harris said. "This technique will allow them to analyze thousands and thousands more scenarios in greater detail than ever before."
Reuters
29-05-2025
- Business
- Reuters
Nvidia, Dell to supply next US Department of Energy supercomputer
BERKELEY, California - May 29 (Reuters) - The U.S. Department of Energy on Thursday said its "Doudna" due in 2026 will use technology from Nvidia and Dell. The computer, named for Nobel Prize-winning scientist Jennifer Doudna who made key CRISPR gene-editing discoveries, will be housed at Lawrence Berkeley National Laboratory in Berkeley, California. At an event at the lab attended by Secretary of Energy Chris Wright, officials said that the system will use Nvidia's latest "Vera Rubin" chips built into liquid-cooled servers by Dell and will be used by 11,000 researchers.
Reuters
20-05-2025
- Business
- Reuters
Google brings AI to grid teams slashing US connection times
May 20 - In April, U.S. power grid operator PJM announced it will deploy Google and Tapestry AI-enhanced tools to accelerate the approval of grid connection requests. PJM operates the largest American power network mostly across eastern states and its partnership with Google is the latest implementation of advanced software by grid operators to improve and speed up connections, data sharing and project management. Google and other power Big Tech groups are major buyers of power and seek faster development of new generation to meet soaring demand from data centers and other computing facilities. Regional Transmission Organisations (RTOs) and Independent System Operators (ISOs) have been inundated with a large number of clean power applications, leading to long delays for project developers. The average length of grid connection studies was 40 months between 2018 and 2022, not including the Texas ERCOT grid, according to a report by Grid Strategies and The Brattle Group in February 2024. Delays differ between markets and ERCOT has independent regulatory status and a more efficient approach. There was 2,600 GW in U.S. grid connection queues at the end of 2023, 95% of it solar, wind and battery storage, according to Lawrence Berkeley Laboratory (Berkeley Lab). This is more than double the total installed generation capacity of 1,280 GW, although not all projects will be completed. Longer timelines impact the speed at which new resources can be brought online, add to costs and contribute to project cancellations, Caitlin Marquis, Managing Director of industry group Advanced Energy United, told Reuters Events. "We find that [connection] costs are rising, that renewable energy projects see significantly higher costs to connect to the grid than fossil fuel projects, and that costs vary widely," Berkeley Lab said in a report in January 2025. "This unpredictability increases the uncertainty and financial risk of renewable development." CHART: US average grid connection costs PJM has already been adopting automation tools to help clear a huge backlog of clean power applications which saw many projects spending more than five years in the connection queue, according to the Federal Energy Regulatory Commission (FERC). Automation has allowed PJM to process about 140 GW out of 200 GW in its queue since 2023, with the remaining 60 GW due for completion by 2026, a PJM spokesperson told Reuters Events. From spring 2026, PJM aims to complete grid connection approvals within one to two years based on impact. Faster connections Inundated with clean power applications, PJM paused all new applications in 2023 and implemented a process to approve projects on a 'first ready, first served' basis, following new rules implemented by the FERC. Previously, projects were addressed on a first come, first served system but this led to vast numbers of speculative projects that were less developed which clogged up grid connection queues. Join hundreds of senior executives across energy, industry and finance at Reuters Events Global Energy Transition 2025. Of the 140 GW projects cleared from the queue since 2023, PJM has signed grid connection agreements with 18 GW and another 63 GW is progressing towards signing agreements over the next two years, although not all of those are guaranteed to proceed, the PJM spokesperson told Reuters Events. From 2026, PJM will require project applications to hit milestones and provide deposits, to reduce speculative applications. CHART: Power generation in grid connection queues at end of 2023 Midwest grid operator MISO and the Southwest Power Pool (SPP) are also using the latest digital technology to slash grid connection times. For several years MISO has used an online submission tool to process grid connection requests and the operator is rolling out a new software tool that will automate and accelerate connection studies. Developed by Pearl Street Technologies, the software tool is "expected to perform the pre-queue and phase 1 studies faster and more efficiently than our existing process which includes file setup and human review," a MISO spokesperson said. In a benchmarking analysis on a $13 billion network upgrade by MISO, the tool completed the Phase 1 study process in just ten days, in comparison with the 686 days using the old system. MISO is now working towards achieving a one-year connection approval process. For exclusive insights on the energy transition, sign up to our newsletter. SPP uses both external and in-house grid connection software 'to find value-add solutions that provide short and long-term advantages in our engineering studies and data management," Derek Wingfield, SPP communications manager told Reuters Events. SPP is also using Pearl Street's software and is also working with Grid Unity, a software company which specialises in grid connection management. Digital technology has enabled SPP to squeeze 'seven years of backlogged [grid connection] requests into three, while also adding two of the three largest clusters on record into the timeframe," Wingfield said. SPP uses an annual portfolio of grid connection agreements to optimise transition planning. Currently, developers are informed of the grid connection costs following completion of studies, but SPP plans to move to a fixed amount upfront charge per megawatt. 'We anticipate this new approach would be a major win for developers to aid their planning during the interconnection process," an SPP spokesperson said. AI future Going forward, grid connection times can be lowered further provided companies are willing to invest and increase data transparency. Many organisations continue to lack internal software engineering expertise or are unwilling to make the upfront financial investment in software programs, Marquis noted. AI can play a pivotal role in accelerating grid connections and increasing the efficiency of grid networks, Ruth Porat, President & Chief Investment Officer, Alphabet and Google, said in a blog post in April. "We are bringing together the power of Alphabet, including Tapestry, Google Cloud and Google DeepMind, to build and deliver a set of collaborative AI tools that will enable PJM to make faster decisions with greater confidence," Porat said. "Creative solutions from across the private and public sectors are crucial to ensure the U.S. has the energy capacity, affordability and reliability needed to capitalize on the opportunity for growth."
