This week, Nvidia announced its next-generation chip, the H200, the latest in Nvidia AI tools, “…the largest chip the worlds ever seen. This is beyond the limits of physics. We had to invent some new technology to make it possible to do this,” Jensen Huang, President of Nvidia.
Designed to manage the stresses of intensive AI model training, the latest prototype will utilize an upgraded memory spec called HBM3e.
Tech companies and military industrial complexes anxiously await its release this coming fall. Large AI language models require thousands of chips to function, each at a hefty cost of $25-40k. The H200 could easily top that figure given it boasts virtually twice the memory capacity of the H100.
“The integration of faster and more extensive HBM memory serves to accelerate performance across computationally demanding tasks including generative AI models and [high-performance computing] applications while optimizing GPU utilization and efficiency,” Ian Buck, Nvidia’s VP of high-performance computing products, said in a video presentation this morning.1
Nvidia AI tools have made significant strides in advancing quantum computing.
Nvidia Quantum Computing
Just a year ago this month, Nvidia announced a new system for accelerated quantum-classical computing. The Nvidia DGX Quantum was the first system to join GPUs and GTC.
Its release marked the introduction of a new platform for “researchers working in high-performance and low-latency quantum-classical computing,”2 allowing them to push the boundaries of innovation with classical, quantum and hybrid algorithm computing methods.
DGX Quantum offers developers access to NVIDIA CUDA Quantum, a robust unified software stack now accessible in open source. This platform, CUDA Quantum, merges quantum and classical computing capabilities, allowing for seamless integration and programming of QPUs, GPUs, and CPUs within a single system.
NVIDIA has unveiled a fresh cohort of partners who are incorporating CUDA Quantum into their platforms. These partners include various quantum hardware companies such as Anyon Systems, Atom Computing, IonQ, ORCA Computing, Oxford Quantum Circuits, and QuEra. Additionally, quantum software companies like Agnostiq and QMware are onboard, alongside supercomputing centers such as the National Institute of Advanced Industrial Science and Technology, the IT Center for Science (CSC), and the National Center for Supercomputing Applications (NCSA).
Boosting Quantum Circuit Simulation:
Financial institutions like investment banks, hedge funds, and asset management firms grapple with significant computational hurdles in conducting large-scale risk simulations and derivative pricing.
Although we recognize the potential of quantum computing in solving financial problems, the reality of physical quantum computers with thousands or millions of qubits remains distant.
However, through the utilization of NVIDIA GPUs and cuQuantum, accelerated quantum circuit simulation becomes achievable. This breakthrough allows for faster experimentation with more qubits and gate depth, empowering financial institutions to stay ahead by developing quantum-accelerated applications.
Combatting Fraud with Quantum Machine Learning:
HSBC, among the world’s largest banks, has pioneered a quantum machine learning application aimed at detecting fraud in digital payments.
Their quantum machine learning algorithm simulated an impressive 165 qubits using NVIDIA GPUs, underscoring NVIDIA’s pivotal role in driving forward quantum research and applications.
Unified Quantum-Classical Computing Platform:
NVIDIA has introduced the Quantum Optimized Device Architecture (QODA), a unified computing platform designed to make quantum computing more accessible. QODA establishes a coherent hybrid quantum-classical programming model, facilitating accelerated breakthroughs in quantum research across diverse domains, including finance.
Fostering Collaboration Across the Quantum Ecosystem:
NVIDIA collaborates closely with partners to develop robust, scalable, and user-friendly tools for quantum-accelerated supercomputing applications. These collaborative efforts empower governments, universities, and industrial corporations to construct practical quantum applications.
In short, NVIDIA assumes a crucial role in advancing quantum computing for financial applications, ranging from accelerating simulations to enabling fraud detection and nurturing collaboration within the quantum ecosystem.
Recently, headquartered in Raleigh, ICO Redhat announced a collaboration with NVIDIA and Fujitsu to enhance and extend the potential for real-time AI data analysis at the edge.
Leveraging technologies developed by the IOWN Global Forum on the foundation of Red Hat OpenShift, the industry’s leading hybrid cloud application platform powered by Kubernetes, this solution has received validation through an IOWN Global Forum’s Proof of Concept.
The application of AI analysis to assess and triage input at the network’s edge promises to reduce latency and increase bandwidth.
