English
Share
Sign In
NVIDIA Announces Roadmap for the Artificial Intelligence Era at Computex 2024
Haebom
The various technologies and vision that NVIDIA showcased at Computex 2024 are not just limited to the announcement of new products, but are expected to become a key driving force that will bring innovation to the entire future industry.
First of all, the 'Physical AI' and 'Omniverse' platforms are expected to open new horizons in the field of robotics. Physical AI is a technology that allows robots to learn and adapt like humans in real environments. Through Omniverse, NVIDIA presented a pipeline that allows these AI models to be sufficiently trained in a virtual environment and applied to actual robots. As realistic physical simulations become possible, it is expected that great progress will be made in the development of industrial robots as well as service robots.
NVIDIA also opened a new chapter in 'digital twin' technology at this event. The Earth-2 project is a digital twin for weather prediction. It is expected to bring about innovation in responding to climate change by combining supercomputing and AI to precisely simulate the Earth's environment. Even in the manufacturing field, digital twins are becoming a key tool for innovation. NVIDIA's Omniverse is expected to become a powerful platform for digital twin implementation throughout the design, simulation, and automation process.
In addition, NVIDIA presented the future of ‘accelerated computing’ at this Computex. In the era of big data and AI, existing general-purpose computing has its limitations. Accelerated computing, which combines the advantages of CPU and GPU, is considered a breakthrough that can solve these problems. In particular, 'Rubin', the next-generation GPU architecture unveiled by NVIDIA, is expected to lead innovation in various fields such as HPC, autonomous driving, and robotics by maximizing computational efficiency.
In addition, NVIDIA also unveiled solutions that can dramatically simplify the development and deployment of AI models, such as 'NIM (NVIDIA Inference Microservice)'. This is a 'game changer' that can change the paradigm of software development, and is expected to greatly contribute to the easier application of AI technology to industrial sites.
In this way, NVIDIA's Computex 2023 announcement is significant in that it presents a vision for future technologies such as artificial intelligence, robotics, and digital twins, while also presenting a comprehensive blueprint that includes specific application plans and the ecosystem. It is expected that it will play a significant role as a bridge connecting cutting-edge technology with industrial innovation.
This is why NVIDIA's future moves are attracting attention. Attention is being paid to whether the vision and strategies revealed this time can lead to concrete results and what changes they will bring to the industrial landscape. This is why we look forward to the leadership and innovation that NVIDIA will show in the fields of artificial intelligence and data science, which will be the core drivers of the 4th Industrial Revolution.
📈
The most important thing is that Rubin, Blackwell's top model, was revealed. I wonder if the era of physical supercomputing is really coming.
Rubin (Ultra): GR100 GPU, R100 SKU, uses HBM4 memory, scheduled for release in 2026-2027.
Blackwell (Ultra): GB200 GPU, B100/B200 SKU, using HBM3e memory, scheduled for release in 2024-2025.
Hopper: GH200/GH100 GPU, H100/H200 SKU, using HBM2e/HBM3/HBM3e memory, scheduled for release in 2022-2024.
Ampere: GA100 GPU, A100 SKU, uses HBM2e memory, available 2020-2022.
Volta: GV100 GPU, V100 SKU, uses HBM2 memory, released in 2018.
Pascal: GP100 GPU, P100 SKU, uses HBM2 memory, released in 2016.
3
/haebom
Subscribe