#Supercomputing

[larimar]

nasaames

[alltimeupdating]

Quantum computers just crossed a milestone — solving problems no traditional machine ever could. From faster drug discovery to smarter traffic systems, the future is now being powered by qubits. 🌍 Explore what this means for healthcare, finance, and our everyday lives.

👉 Read the full story and see how this breakthrough changes everything.

[impact-newswire]

NVIDIA Brings AI Supercomputer Manufacturing to the U.S. for the First Time

NVIDIA Blackwell chip production starts in Arizona — NVIDIA opens first US factories. NVIDIA is working with its manufacturing partners to design and build factories that, for the first time, will produce NVIDIA AI supercomputers entirely in the U.S. Together with leading manufacturing partners, the company has commissioned more than a million square feet of manufacturing space to build and…

[jade-malay]

Dallas Innovation: Where AI Shapes the Future

AI and supercomputing aren't just the future —they're the new foundation of innovation right here in Dallas! Where cutting-edge technology, top talent, and bold vision come together to build a smarter, brighter world. #JadeMalay #InnovationHub #DallasTech #FutureReady

[newslink7com]

Nvidia Hit With $5.5 Billion Loss After U.S. Government Bans H20 AI Chip Exports to China Over National Security and Supercomputing Risks

The U.S. just blocked Nvidia’s most advanced AI chip for China, citing fears of supercomputing misuse. The result? A staggering $5.5B charge and global tech ripple effects.

👉 Read the full story at NewsLink7.com

[newstodays1]

Nanotechnology and Supercomputing: A Synergy for the Future

Introduction The intersection of nanotechnology and supercomputing is driving unprecedented advancements in computational power, energy efficiency, and miniaturization. As traditional silicon-based computing approaches its physical limits, nanomaterials and quantum-scale innovations are paving the way for next-generation supercomputers capable of solving complex global challenges. This blog…

[sentivium]

Project Digits: How NVIDIA's $3,000 AI Supercomputer Could Democratize Local AI Development | Caveman Press

[karannnnn69]

Supercomputing: The Key to National Leadership

[ai-network]

Elon Musk is Breaking the GPU Coherence Barrier

In a significant development for artificial intelligence, Elon Musk and xAI has reportedly achieved what experts deemed impossible: creating a supercomputer cluster that maintains coherence across more than 100,000 GPUs. This breakthrough, confirmed by NVIDIA CEO Jensen Huang as "superhuman," could revolutionize AI development and capabilities. The Challenge of Coherence Industry experts previously believed it was impossible to maintain coherence—the ability for GPUs to effectively communicate with each other—beyond 25,000-30,000 GPUs. This limitation was seen as a major bottleneck in scaling AI systems. However, Musk's team at xAI has shattered this barrier using an unexpected solution: ethernet technology. The Technical Innovation xAI's supercomputer, dubbed "Colossus," employs a unique networking approach where each graphics card has a dedicated 400GB network interface controller, enabling communication speeds of 3.6 terabits per second per server. Surprisingly, the system uses standard ethernet rather than the exotic connections typically found in supercomputers, possibly drawing from Tesla's experience with ethernet implementations in vehicles like the Cybertruck. Real-World Impact Early evidence of the breakthrough's potential can be seen in Tesla's Full Self-Driving Version 13, which reportedly shows significant improvements over previous versions. The true test will come with the release of Grok 3, xAI's next-generation AI model, expected in January or February. Future Implications The team plans to scale the system to 200,000 GPUs and eventually to one million, potentially enabling unprecedented AI capabilities. This scaling could lead to: More intelligent AI systems with higher "IQ" levels Better real-time understanding of current events through X (formerly Twitter) data integration Improved problem-solving capabilities in complex fields like physics The Investment Race and the "Elon Musk Effect" This breakthrough has triggered what experts call a "prisoner's dilemma" in the AI industry. Major tech companies now face pressure to invest in similar large-scale computing infrastructure, with potential investments reaching hundreds of billions of dollars. The stakes are enormous—whoever achieves artificial super intelligence first could create hundreds of trillions of dollars in value. This development marks another instance of "Elon Musk Effect" in which Musk's companies continue to defy industry expectations, though it's important to note that while Musk is credited with the initial concept, the implementation required the effort of hundreds of engineers. The success of this approach could reshape the future of AI development and computing architecture. Read the full article

[govindhtech]

What Is Quantum Centric Supercomputing? And How It Works

What is Quantum centric supercomputing?

In order to develop a computing system that can tackle very complicated real-world issues, quantum centric supercomputing, a groundbreaking approach to computer science, blends quantum computing with conventional high-performance computing (HPC).

Using error mitigation and error correction methods, a quantum-centric supercomputer is a next-generation combination of a quantum computer with a classical supercomputer that produces results in real-world runtimes.

It is anticipated that in the age of quantum computing, quantum-centric supercomputing would enable scientists to make significant advances in generative AI, postquantum cryptography, machine learning, material sciences, and other areas, maybe even surpassing large-scale fully quantum systems.

A fully functional quantum-centric supercomputer integrates quantum circuitry with traditional computing resources through sophisticated middleware. The fundamental components of quantum centric supercomputing, which are based on the IBM Quantum System Two architecture, integrate quantum technology with conventional supercomputers to enhance and complement their respective capabilities.

How Quantum centric supercomputing work?

The quantum processing unit (QPU) is the central component of a quantum centric supercomputing. IBM’s QPU consists of a multilayer semiconductor chip etched with superconducting circuits and the gear that receives and outputs circuits. These circuits house the qubits that are utilized for computations as well as the gates that manipulate them. The circuits are separated into many layers of input and output wire, a layer with resonators for readout, and a layer containing the qubits. Interconnects, amplifiers, and signal-filtering components are also included in the QPU.

A superconducting capacitor connected to elements known as Josephson junctions, which function similarly to lossless, nonlinear inductors, makes up the kind of physical qubit that IBM uses. Only certain values may be assumed by the current flowing across Josephson junctions due to the superconducting nature of the system. Additionally, only two of those particular values are available due to the Josephson junctions spacing them away.

The lowest two current values zero and one, or a superposition of both zero and one are then used to encode the qubit. Programmers use quantum instructions, often referred to as gates, to couple qubits together and alter their states. These are a number of microwave waveforms that have been particularly created.

Some of the QPU components must be kept within a dilution refrigerator, which uses liquid-helium to keep them cool, in order to maintain the qubits’ proper working temperature. Classical computing hardware at normal temperature is needed for other QPU components. The QPU is then linked to runtime infrastructure, which handles results processing and error mitigation. This computer is quantum.

By enabling smooth communication between the two, middleware and hybrid cloud solutions enable the integration of quantum and classical systems. Without requiring a total redesign of present infrastructures, this hybrid technique helps guarantee that quantum processing units may be utilized efficiently within quantum computers coupled to conventional computing frameworks, optimizing their impact.

Quantum centric supercomputing use cases

Large-scale data processing might be accelerated by quantum computers, which are particularly good at tackling some challenging issues. Quantum computing may provide the key to advancements in a number of crucial fields, including material research, supply chain optimization, medication development, and climate change issues.

Pharmaceuticals: Research and development of novel, life-saving medications and medical treatments can be greatly accelerated by quantum computers that can simulate molecular behavior and biochemical interactions.

Chemistry: Quantum computers may influence medical research for the same reasons, but they may also offer previously unidentified ways to reduce hazardous or damaging chemical byproducts. Better procedures for the carbon breakdown required to tackle climate-threatening emissions or better catalysts that enable petrochemical alternatives can result from quantum computing.

Machine learning: Researchers are investigating whether some quantum algorithms would be able to see datasets in a novel way, offering a speedup for specific machine learning tasks, as interest and investment in artificial intelligence (AI) and related disciplines like machine learning increase.

Challenges Of Quantum centric supercomputing

Today’s quantum computers are scientific instruments that can execute some programs more effectively than conventional simulations, at least when modeling particular quantum systems. Nonetheless, quantum computing will continue to be beneficial for the foreseeable future when combined with current and upcoming conventional supercomputing. As a result, quantum scientists are getting ready for a time when quantum circuits will be able to assist traditional supercomputers in solving issues.

The development of the middleware that enables communication between classical and quantum computers, as well as general issues with quantum computers themselves, are the main obstacles facing quantum centric supercomputing. The following major challenges have been recognized by developers to be addressed prior to attaining quantum advantage.

Enhancing Interconnects

Millions of physical qubits are needed to create a fully functional large-scale quantum computer. However, scaling individual chips to these levels is extremely difficult due to real hardware limits. IBM is creating next-generation interconnects that can transfer quantum information between many devices as a remedy. To achieve the necessary qubits for error correction, this method offers modular scalability.

IBM intends to use proof-of-concept chips dubbed Flamingo and Crossbill, respectively, to show these novel interconnects, which are referred to as l-couples and m-couplers. Chip scaling is the responsibility of these couplers. IBM intends to use a chip known as Kookaburra to demonstrate c-couplers by the end of 2026. They are in charge of helping to fix errors.

Scaling quantum processors

Current quantum processors can only handle a small number of possible qubits, despite the fact that quantum processors based on qubits utilized in quantum computing have the potential to significantly surpass bit-based processors. IBM intends to launch a quantum system with 200 logical qubits that can execute 100 million quantum gates by 2029 as research advances, with a target of 2,000 logical qubits that can execute 1 billion gates by 2033.

Scaling quantum hardware

Qubits require massive cooling systems that can produce temperatures lower than space since, despite their power, they are also very prone to errors. In order to lower footprint, cost, and energy consumption, researchers are creating methods to scale qubits, electronics, infrastructure, and software.

Quantum error correction

Although qubit coherence is fleeting, it is essential for producing precise quantum data. One of the biggest challenges for any quantum system is decoherence, which is the process by which qubits malfunction and provide erroneous outputs. Encoding quantum information into more qubits than would otherwise be necessary is necessary for quantum error correction. IBM unveiled a revolutionary new error-correcting code in 2024 that is around ten times more effective than previous techniques. This new code paves the way for the operation of quantum circuits with a billion logic gates or more, even if error correction is still an open subject.

Quantum algorithm discovery

Two elements are necessary for quantum advantage. The first consists of feasible quantum circuits, and the second is a technique to show that, in comparison to other state-of-the-art approaches, such quantum circuits are the most effective way to tackle a quantum issue. Current quantum technologies will go from quantum usefulness to quantum advantage with the discovery of quantum algorithms.

Quantum software and middleware

In order to design, optimize, and run quantum programs, the core of quantum algorithm discovery depends on an extremely reliable and powerful software stack. By far the most used quantum software in the world is IBM’s Qiskit. Its open source SDK and related tools and services are built on Python and may be used to execute on IBM’s fleet of superconducting quantum computers as well as on systems that employ other technologies, such quantum annealing or ions trapped in magnetic fields.

Read more on govindhtech.com

[winbuzzer]

ICYMI: In an escalation of its legal defense over supercomputing patents, Munich-based ParTec AG has sued Nvidia for allegedly infringing upon its intellectual property essential to AI-focused supercomputing. #Nvidia http://dlvr.it/TFt9Fb

[knowledge-wale]

NVIDIA's Grace Hopper Superchip is revolutionizing AI workloads in data centers, designed specifically to handle massive AI and machine learning tasks. It combines a 144-core ARM CPU with a powerful GPU, delivering up to 10x faster performance for AI training and inference compared to traditional systems. The chip is optimized for H100 Tensor Core GPUs, pushing the boundaries of data processing speeds and reducing latency, making it ideal for next-gen AI applications like self-driving cars, language models, and real-time analytics. This breakthrough is expected to accelerate the AI industry's evolution and drive more efficient data center architectures. https://t.ly/2jokn

[retrocompmx]

Un 28 de septiembre, pero de 1925, nace Seymour Cray quien es conocido como el "padre de la supercomputación".

Fundador de Cray Research, revolucionó la informática con la creación de las primeras supercomputadoras de alto rendimiento.

Su diseño pionero del Cray-1 en 1976 estableció nuevos estándares de velocidad y potencia, permitiendo avances científicos y tecnológicos en todo el mundo.

A lo largo de su carrera, Cray se destacó por su enfoque innovador en la arquitectura de computadoras y la refrigeración líquida para maximizar el rendimiento, así como la invención del procesador RISC.

La idea de usar transistores de en lugar de tubos al vacío, es su legado. Un legado que sigue vivo en la industria de la computación de alto nivel.

#retrocomputingmx#seymourcray#Supercomputación#historiadelacomputación

[ourwitching]

CUDA Graphs can provide a significant performance increase, as the driver is able to optimize exe...