Perspective: Nvidia and Microsoft are working on a virtual supercomputer with GPU-based Azure instances. The goal of the development is to accelerate the latest AI algorithms to create even more realistic works of art or conduct AI research.
Generative AI models are proving useful for many applications. Machine learning algorithms can create supernatural images or predict source code from the future, often negatively influencing public opinion with their ability to “fall into the wrong hands.” A new partnership between the two biggest tech companies promises to accelerate these opportunities by building an “AI supercomputer” in the cloud.
Nvidia and Microsoft have announced a “multi-year collaboration”. build the world’s most powerful supercomputer specifically designed to accelerate artificial intelligence and machine learning algorithms. The partnership uses the Microsoft Azure cloud platform combined with high performance Nvidia GPU. Several other components will speed up the entire communications stack.
Hardware includes “tens of thousands” of Nvidia A100 (Ampere-based) and H100 (Hopper) enterprise GPUs. The cloud infrastructure includes GPU-accelerated Microsoft ND and NC series virtual machines. Quantum-2 400 Gb/s InfiniBand network technology and AI Enterprise software package connect the communication device.
In essence, this is a new design of an artificial intelligence supercomputer. functions as a cloud service that works with Azure instances. Nvidia explained that customers can purchase resources “as usual with a real supercomputer” while the software layer reserves the necessary virtual machines.
It is architecturally the same as a physical supercomputer, but runs on virtual machines “in the cloud”. The most obvious advantage is that it does not require a special (and massive) physical device in the research laboratory. However, the capabilities provided will allow enterprises to scale virtual instances “up to the status of a supercomputer.”
The main focus of the virtual supercomputer is on the improvement and development of generative AI models. This rapidly growing area of AI research is currently relies on models such as the Megatron Turing NLG 530B as the basis for “unsupervised self-learning algorithms to generate new text, code, digital images, video or audio”.
Nvidia highlighted how AI technology is accelerating, industry adoption is growing, and breakthroughs in the field have sparked a tidal wave of research, new startups, and enterprise applications. The partnership with Microsoft will provide clients and researchers with state-of-the-art AI infrastructure and software to capitalize on the transformative power of AI.
Or, in Microsoft’s words, powerful AI capabilities “for every enterprise in Microsoft Azure.”