News/Computing
Nvidia partners with iGenius to build AI supercomputer
The collaboration between Italian AI startup iGenius and tech giant Nvidia marks a significant advancement in European AI infrastructure with the unveiling of the Colosseum supercomputer. Project overview: The Colosseum supercomputer represents a major investment in European AI capabilities, specifically designed to support advanced AI models in highly regulated industries. The system delivers 115 exaflops of computational power through Nvidia Grace Blackwell Superchips The infrastructure will support training of open-source generative AI and large language models exceeding one trillion parameters The system achieves 25 times greater energy efficiency compared to previous Nvidia computing platforms Technical infrastructure: iGenius is establishing a...
read Dec 9, 2024IBM’s new tech enables energy-efficient computing with light-speed data transmission
Breakthrough Innovation: IBM has unveiled new optical technology that enables data transmission at the speed of light within data centers, marking a significant advance in computing infrastructure. The technology utilizes polymer optical wave guides, which combine electrical and optical connectivity in their circuit design These advanced optics can operate in extreme conditions, functioning in temperatures from -40°C to 125°C and in humid environments The system demonstrates remarkable durability and versatility for data center applications Environmental Impact: The new optical technology promises substantial reductions in energy consumption and environmental impact for AI operations. Each AI model trained using this technology could...
read Dec 7, 2024Quantum AI robots that match human intelligence are coming, researchers say
The intersection of quantum computing and artificial intelligence promises to usher in a new era of advanced robotics with capabilities approaching human-level intelligence. Key breakthrough: Scientists have proposed combining quantum computing with AI to create "qubots" - robots that could potentially match human intelligence by leveraging quantum algorithms and processes. These advanced robots would overcome the binary computing limitations that currently restrict robotic capabilities Qubots would excel at processing vast amounts of sensory data and providing real-time responses The technology could enable sophisticated navigation, decision-making, and multi-robot coordination Technical advantages: Quantum computing's superior processing power offers unprecedented potential for advancing...
read Dec 6, 2024Understanding quantum AI and how it will reshape the world
The intersection of quantum computing and artificial intelligence represents a significant technological frontier that promises to enhance computational capabilities and transform various industries. The foundations of Quantum AI: Quantum AI merges quantum mechanics principles with artificial intelligence algorithms to create more powerful computing solutions that transcend traditional computational limits. Quantum computers utilize qubits instead of traditional binary bits, allowing them to process information in multiple states simultaneously This fundamental difference enables quantum systems to handle complex calculations and data processing tasks with unprecedented efficiency The technology represents a significant departure from conventional computing architectures that rely on binary (0 or...
read Dec 5, 2024Los Alamos and U-Michigan join forces for AI research
The U.S. Department of Energy's Los Alamos National Laboratory is partnering with the University of Michigan to establish new artificial intelligence research facilities, marking a significant expansion of their long-standing collaboration. Partnership Overview: The initiative involves creating two distinct computing centers near Ypsilanti, Michigan, focused on artificial intelligence and high-performance computing research. The 20-acre property at 10221 Textile Road will host both classified and non-classified research facilities A five-year, $15-million research contract was established earlier this year to develop advanced technologies and address clean energy challenges Funding for the centers will come from federal and state economic development sources Facility...
read Dec 3, 2024Global AI computing to require eight New York’s worth of power by 2026
The expansion of artificial intelligence computing infrastructure is creating unprecedented demands on power grids, with projections showing dramatic increases in energy consumption over the next few years. Current state of AI computing: The artificial intelligence industry is entering a new phase where the focus is shifting from model training to widespread deployment of AI systems in production environments. AI inference workloads, which involve running trained models in real-world applications, are becoming a major driver of computing demand At least twelve new AI data centers are currently planned or under construction, each requiring approximately one gigawatt of power to operate The...
read Dec 3, 2024Amazon is building the world’s biggest AI supercomputer with Anthropic
Amazon has partnered with Anthropic to construct one of the world's most powerful AI supercomputers. Project scope and specifications: Amazon's new supercomputer project, dubbed Rainer, represents a major advancement in AI computing infrastructure and will leverage hundreds of thousands of Amazon's latest Trainium 2 chips. The system will be five times larger than Anthropic's current model-building cluster When completed, it is expected to be the world's largest reported AI machine Amazon has invested $8 billion in Anthropic, positioning itself as a major player in the AI infrastructure space Technical innovations: Amazon's new chip developments and infrastructure improvements signal a significant...
read Dec 3, 2024VC icon Steve Jurvetson shares insights on Moore’s Law, future of AI
The rapid evolution of computing power and its implications for technology and society can be traced back much further than commonly believed, with significant insights from venture capitalist Steve Jurvetson highlighting the broader historical context of technological advancement. Historical context and origins: Moore's Law, traditionally associated with transistor density, actually began its trajectory in the 1800s with Babbage's analytical engine. The computational growth pattern predates Gordon Moore's famous prediction about transistors Early computing milestones, including Hollerith cards, were part of this longer technological progression Jurvetson suggests that Moore's Law represents a broader trend in computational capability rather than just transistor...
read Dec 2, 2024Nous Research trains AI model with global distributed computing
The development of distributed AI training methods marks a significant shift in how large language models can be created, potentially democratizing access to AI development beyond major tech companies and specialized data centers. Key breakthrough: Nous Research is pre-training a 15-billion parameter large language model using machines distributed across the internet, departing from traditional centralized data center approaches. The training process is being livestreamed on distro.nousresearch.com, showing real-time evaluation benchmarks and hardware locations across the U.S. and Europe The project utilizes Nous DisTrO (Distributed Training Over-the-Internet), reducing inter-GPU communication bandwidth requirements by up to 10,000x The system can operate on...
read Dec 2, 2024Wearable tech and the computing paradigm of the future
The relationship between generative AI and wearable computing is emerging as a critical pathway in the evolution of human-computer interaction, potentially marking a significant shift in how we interact with technology. The evolution of computing paradigms: The progression of computing has followed distinct stages, from mainframes to personal computers to smartphones, with each new paradigm building upon the foundations laid by its predecessors. Each major computing shift has been characterized by fundamental changes in how users interact with devices, from punch cards to keyboards to touchscreens The transition between paradigms has historically been facilitated by applications that bridge the gap...
read Dec 2, 2024Nebius secures funding to address the AI industry’s computing needs
Artificial intelligence is creating unprecedented demand for specialized computing infrastructure, particularly the GPU-powered cloud platforms needed to handle intensive AI workloads. The foundation of AI innovation: Nebius has emerged as a full-stack AI-native platform designed specifically for handling demanding artificial intelligence and machine learning computational tasks. Built by a team of experienced hardware and software engineers, Nebius offers a vertically integrated solution that spans from data center operations to cloud platform services. The company brings significant technical expertise from years of building large-scale infrastructure systems. With approximately 400 engineers globally, Nebius combines substantial resources with an entrepreneurial mindset. Leadership and...
read Dec 2, 2024MIT-developed photonic chip promises fast AI computations with less energy
Photonic computing has emerged as a promising solution to address the growing computational demands of complex machine learning models, with researchers developing a breakthrough chip that processes neural network operations using light instead of electricity. Key Innovation: MIT researchers and collaborators have created a fully integrated photonic processor that performs all essential deep neural network computations optically on a single chip, achieving remarkable speed and efficiency. The chip completed machine learning classification tasks in under half a nanosecond while maintaining 92% accuracy, matching traditional electronic hardware performance Built using commercial foundry processes, the technology shows potential for scalable manufacturing and...
read Dec 1, 2024Epoch’s new simulator offers visualizations of real-time and historical AI training scenarios
The release of Epoch AI's Distributed Training Interactive Simulator marks a significant advancement in understanding and optimizing large language model training configurations. Core functionality: The simulator enables detailed modeling of distributed training runs for large language models, incorporating bandwidth and latency costs across GPU clusters. The platform provides real-time visualization through training FLOP versus model FLOP utilization plots Users can toggle between preset configurations or create custom scenarios to explore different training parameters The tool accounts for critical variables including dataset size, batch size, model depth, and GPU specifications Technical capabilities: The simulator's comprehensive approach to modeling distributed training encompasses...
read Dec 1, 2024AI and quantum tech to create 1 million jobs in India by 2030
The rapid expansion of emerging technologies in India's IT sector is poised to create significant employment opportunities over the next six years, with particular growth in cybersecurity, artificial intelligence, and quantum computing. Market growth and job creation: Generative AI and quantum computing are projected to generate more than one million jobs in India by 2030, according to a recent Quess Corp report. Cybersecurity roles have seen a remarkable 58% growth in Q2FY25 DevOps positions have increased by 25% during the same period Development roles currently constitute 40% of all tech hiring AI and machine learning positions have grown by 30%...
read Nov 30, 2024The Big Tech companies buying the most GPUs
The rapid expansion of AI computing infrastructure among major technology companies is reshaping the competitive landscape of artificial intelligence development and deployment. Current computing landscape: The distribution of high-performance AI chips, particularly Nvidia's H100 GPUs and equivalent processors, reveals significant disparities among leading tech companies in their AI computing capabilities. Google leads the pack with an estimated 1-1.5 million H100-equivalent chips by the end of 2024, combining both Nvidia GPUs and their custom TPU processors Microsoft follows with 750,000-900,000 units, reflecting their strategic partnership with OpenAI and aggressive AI infrastructure investments Meta's projected 550,000-650,000 chips aligns with their ambitious AI...
read Nov 29, 2024The economics of LLM operations every business leader should know
Market dynamics overview: The enterprise AI market is experiencing a dramatic decrease in the cost of LLM operations, measured in tokens (the basic units of text that AI models process). The cost of LLM performance is declining approximately 10x annually, making advanced AI capabilities increasingly accessible to businesses This price reduction is driven by smaller models, open-source developments, and improved optimization techniques Companies like OpenAI, Meta, Google, and Anthropic are competing to deliver better performance at lower costs Technical benchmarks and measurements: Performance evaluation relies on standardized testing methods to assess LLM capabilities across various domains. The MMLU (Measuring Massive...
read Nov 27, 2024AMD launches ROCm 6.3, an open-source platform to reduce compute costs
The AMD ROCm Version 6.3 release marks a significant advancement in open-source software for AI, machine learning, and high-performance computing on AMD Instinct GPU accelerators. Major updates and core features: ROCm 6.3 introduces several key improvements aimed at enhancing developer productivity and computational performance across various sectors. SGLang integration enables up to 6X higher performance for Large Language Model (LLM) inferencing A re-engineered FlashAttention-2 implementation provides up to 3X faster processing for AI model training New multi-node Fast Fourier Transform (FFT) capabilities support distributed computing applications Enhanced computer vision libraries include support for AV1 codec and improved JPEG processing AI...
read Nov 26, 2024MediaTek’s new processors will power on-device Agentic AI in smartphones
The MediaTek Dimensity 9400 represents a significant advancement in mobile processor technology, promising to power the next generation of flagship smartphones with enhanced performance, AI capabilities, and power efficiency. Core architecture and performance: MediaTek's 2nd Gen All Big Core design marks a significant evolution in mobile processing architecture. Built on Arm's v9.2 CPU architecture, the chip combines one Cortex-X925 core, three Cortex-X4 cores, and four Cortex-A720 cores The processor delivers up to 35% faster single-core and 28% faster multi-core performance compared to its predecessor Support for LPDDR5X 10667 memory enables 10.7Gbps bandwidth for faster data transfer and improved multitasking AI...
read Nov 22, 2024How AI is rewriting the rules of enterprise edge computing
The rapid adoption of artificial intelligence applications in enterprise environments is fundamentally changing the requirements for edge network infrastructure, particularly in how organizations handle data traffic and network resources. The AI networking challenge: Enterprise networks are facing unprecedented demands from AI applications that introduce radically different traffic patterns and resource requirements compared to traditional web applications. AI workloads generate bursty, unpredictable traffic that requires symmetrical upload and download capabilities Traditional edge networks, designed primarily for downstream-heavy web traffic, struggle to handle AI's unique demands Applications like generative AI and video inferencing require significantly higher bandwidth and lower latency than conventional...
read Nov 22, 2024What does the future hold for WebGPU?
The WebGPU specification continues to evolve through collaborative efforts of major tech companies, with significant developments emerging from recent GPU for the Web working group meetings. Standardization progress: The specification is moving closer to achieving W3C candidate recommendation status, marking a crucial step toward broader implementation and stability. Meeting participants agreed there are no major blockers preventing the achievement of Milestone 0 The transition to candidate recommendation status will provide stronger guarantees of stability and intellectual property protection Stakeholders expressed confidence that remaining issues can be resolved efficiently AI-focused enhancements: Several key features are being prioritized to improve WebGPU's artificial...
read Nov 21, 2024AI will drive major scientific advances, NVIDIA CEO tells SC24
The intersection of artificial intelligence and scientific computing is reaching new heights as NVIDIA unveils groundbreaking tools and technologies at the SC24 conference in Atlanta. Strategic vision and historical context: NVIDIA's 25-year journey from creating the first GPU to pioneering AI-driven scientific computing represents a fundamental shift in computational capabilities. CEO Jensen Huang emphasized how supercomputers have become essential instruments for scientific discovery The company's CUDA platform, introduced in 2006, has reduced computing costs by a factor of one million Key milestones include the Tsubame supercomputer (2008), Oak Ridge's Titan (2012), and the AI-focused DGX-1 (2016) Core technological advancements: NVIDIA...
read Nov 21, 2024Google’s AlphaQubit addresses one of quantum computing’s biggest hurdles
The development of AlphaQubit by Google DeepMind and Quantum AI teams marks a significant advancement in addressing quantum computing's persistent error correction challenges, potentially bringing us closer to practical quantum computers. Key breakthrough: Google DeepMind and Quantum AI teams have created AlphaQubit, an artificial intelligence decoder that identifies quantum computing errors with unprecedented accuracy. The system uses a neural network based on Transformer architecture to detect when logical qubits deviate from their intended state Testing on a 49-qubit Sycamore quantum processor and simulated systems up to 241 qubits demonstrated AlphaQubit's capabilities The AI-powered system showed remarkable adaptability, maintaining performance even...
read Nov 19, 2024How to unlock the potential of mobile artificial intelligence
Mobile artificial intelligence is emerging as a critical frontier in technology, with organizations seeking ways to bring sophisticated AI capabilities directly to smartphones and Internet of Things (IoT) devices rather than relying solely on cloud computing. Core technical challenge: The fundamental hurdle in mobile AI deployment stems from the significant gap between the computational demands of AI systems and the limited processing power available on mobile devices. Mobile devices typically possess only a fraction of the computing resources found in cloud data centers Running complex AI models locally requires careful optimization and architectural planning Edge computing, which processes data closer...
read Nov 18, 2024Napkin math: Rapidly declining inference costs are making AI assistants cheaper than ever
The rapid decline in AI processing costs is making personal AI assistants, similar to those depicted in science fiction, increasingly feasible from a cost perspective. Current cost analysis: The computational cost for an AI assistant that interacts with users as frequently as they check their smartphones has dropped to remarkably low levels. Based on average smartphone usage patterns of 144 daily interactions, with four exchanges per interaction over 30 days, the raw computing cost would be approximately 75 cents per month This calculation assumes processing about 0.5 million tokens monthly at a rate of 15 cents per million tokens Even...
read