New “exascale” supercomputers will bring breakthroughs in science. But the technology also exists to study nuclear weapons.
Category: supercomputing – Page 42
Researchers from the University of Sussex and Universal Quantum have demonstrated for the first time that quantum bits (qubits) can directly transfer between quantum computer microchips and demonstrated this with record-breaking speed and accuracy. This breakthrough resolves a major challenge in building quantum computers large and powerful enough to tackle complex problems that are of critical importance to society.
Today, quantum computers operate on the 100-qubit scale. Experts anticipate millions of qubits are required to solve important problems that are out of reach of today’s most powerful supercomputers. There is a global quantum race to develop quantum computers that can help in many important societal challenges from drug discovery to making fertilizer production more energy efficient and solving important problems in nearly every industry, ranging from aeronautics to the financial sector.
In the research paper, published today in Nature Communications, the scientists demonstrate how they have used a new and powerful technique, which they dub “UQ Connect,” to use electric field links to enable qubits to move from one quantum computing microchip module to another with unprecedented speed and precision. This allows chips to slot together like a jigsaw puzzle to make a more powerful quantum computer.
This is still the beginning of what AI can possibly do.
IBM’s Watson supercomputer is working wonders in an area where OpenAI’s ChatGPT does not have much to offer, the stock market. An exchange-traded fund (ETF) is using the power of artificial intelligence (AI) to balance its portfolio and has done pretty well for itself this year, ETF.
PhonlamaiPhoto/iStock.
ChatGPT responded that the stock market was too hard to predict and that it did not have access to live stock data. However, ETF Managers Group, in partnership with a fintech firm Equbot has been using AI to pick holdings in the $102 million AI-powered Equity ETF (AIEQ) since 2017. The fund has doubled the returns on the Vanguard Total Stock Market ETF (VTI) this year.
Physicists have invented a new type of analog quantum computer that can tackle hard physics problems that the most powerful digital supercomputers cannot solve.
New research published in Nature Physics by collaborating scientists from Stanford University in the U.S. and University College Dublin (UCD) in Ireland has shown that a novel type of highly-specialized analog computer, whose circuits feature quantum components, can solve problems from the cutting edge of quantum physics that were previously beyond reach. When scaled up, such devices may be able to shed light on some of the most important unsolved problems in physics.
For example, scientists and engineers have long wanted to gain a better understanding of superconductivity, because existing superconducting materials —such as those used in MRI machines, high speed train and long-distance energy-efficient power networks—currently operate only at extremely low temperatures, limiting their wider use. The holy grail of materials science is to find materials that are superconducting at room temperature, which would revolutionize their use in a host of technologies.
Tech giants from Google to Amazon and Alibaba —not to mention nation-states vying for technological supremacy—are racing to dominate this space. The global quantum-computing industry is projected to grow from $412 million in 2020 to $8.6 billion in 2027, according to an International Data Corp. analysis.
Whereas traditional computers rely on binary “bits”—switches either on or off, denoted as 1s and 0s—to process information, the “qubits” that underpin quantum computing are tiny subatomic particles that can exist in some percentage of both states simultaneously, rather like a coin spinning in midair. This leap from dual to multivariate processing exponentially boosts computing power. Complex problems that currently take the most powerful supercomputer several years could potentially be solved in seconds. Future quantum computers could open hitherto unfathomable frontiers in mathematics and science, helping to solve existential challenges like climate change and food security. A flurry of recent breakthroughs and government investment means we now sit on the cusp of a quantum revolution. “I believe we will do more in the next five years in quantum innovation than we did in the last 30,” says Gambetta.
But any disrupter comes with risks, and quantum has become a national-security migraine. Its problem-solving capacity will soon render all existing cryptography obsolete, jeopardizing communications, financial transactions, and even military defenses. “People describe quantum as a new space race,” says Dan O’Shea, operations manager for Inside Quantum Technology, an industry publication. In October, U.S. President Joe Biden toured IBM’s quantum data center in Poughkeepsie, N.Y., calling quantum “vital to our economy and equally important to our national security.” In this new era of great-power competition, China and the U.S. are particularly hell-bent on conquering the technology lest they lose vital ground. “This technology is going to be the next industrial revolution,” says Tony Uttley, president and COO for Quantinuum, a Colorado-based firm that offers commercial quantum applications. “It’s like the beginning of the internet, or the beginning of classical computing.”
The expertise of GPT3.5 at the industrial scale.
If you are tired of your requests to access ChatGPT being waitlisted repeatedly, Microsoft has some good news for you. The chatbot is coming soon to Azure Open AI services, where businesses can access the most advanced artificial intelligence (AI) in the world, the company said in a press release.
ChatGPT, the chatbot released on November 30 last year, has caught the imagination of engineers and non-engineers alike. The large language model used by the platform allows the AI to help answer user queries in a conversational style.
NurPhoto/Getty.
Microsoft teamed up with OpenAI in July 2019 to accelerate breakthroughs in the field of AI. On its part, Microsoft used its expertise in computing to build AI supercomputers exclusively for OpenAI and, since November 2021, has been offering the Azure OpenAI service for enterprise customers.
My first computer had a CPU with 3,510 transistors. We now live in a world where you can get chips with over a trillion transistors.
“This a data centre accelerator that contains 146 billion transistors.”
I checked and the article didn’t include the transistors that made up the L4 cache memory on the chip. The actual total is 1.25 trillion transistors plus another 1.1 trillion capacitors.
This chip is coming out later this year and I expect to see it used in large quantities in supercomputers and in the server market in general.
Quite impressive!
By the way, the record for transistors on a chip is 2.6 trillion with the latest Cerebras chip and that chip is two years old and doesn’t even use any chiplets, so is just one giant chip. (Compared to the MI300 composed of 41 chiplets!)
Quantum computers hold the promise of performing certain tasks that are intractable even on the world’s most powerful supercomputers. In the future, scientists anticipate using quantum computing to emulate materials systems, simulate quantum chemistry, and optimize hard tasks, with impacts potentially spanning finance to pharmaceuticals.
However, realizing this promise requires resilient and extensible hardware. One challenge in building a large-scale quantum computer is that researchers must find an effective way to interconnect quantum information nodes—smaller-scale processing nodes separated across a computer chip. Because quantum computers are fundamentally different from classical computers, conventional techniques used to communicate electronic information do not directly translate to quantum devices. However, one requirement is certain: Whether via a classical or a quantum interconnect, the carried information must be transmitted and received.
To this end, MIT researchers have developed a quantum computing architecture that will enable extensible, high-fidelity communication between superconducting quantum processors. In work published in Nature Physics, MIT researchers demonstrate step one, the deterministic emission of single photons—information carriers—in a user-specified direction. Their method ensures quantum information flows in the correct direction more than 96 percent of the time.
The Human Digital Twin
Posted in biotech/medical, health, supercomputing
We are living a data revolution in the biomedical field, and scientific research is advancing at an unprecedented speed to improve modern medicine. One of the key aspects of such medicine is the tailoring of treatments to each patient, by analising the specific changes that led to disease along with the unique characteristics with which the person was born. The use of supercomputers is essential to make sense of the vast amounts of data, and to simulate aspects of our bodies to calculate for instance which drug is more appropriate for each patient for a given disease.
This video showcases some of the research done at the Life Sciences Department of the Barcelona Supercomputing Center, contributing to a better understanding of our bodies in health and disease, and to a future where a Human Digital Twin can help to live healthier and longer.
The race is on to develop a quantum computer that can outpace a conventional supercomputer, and researchers from around the world are full-steam ahead. If scaled to adequate sizes, quantum computers represent the largest leap forward in computing for decades, carrying the potential to leave our current machines in the dust, but significant hurdles still remain.
Now, a team of researchers from China have created a superconducting quantum processor with 66 functional qubits which, when faced with a complex sampling task, was able to blast past even the most powerful supercomputers and complete it in just a fraction of the time. What makes the research so impressive is how it demonstrates a huge leap towards quantum primacy, a milestone in which quantum computers complete a task that is infeasible for a conventional computer to complete.