We all know the role that NVIDIA plays in the graphics card market. Probably many users also know that this company has been leading the market for artificial intelligence chips for years that reside in the servers on which ChatGPT runs, among many other services. In fact, it currently monopolizes about 80% of this industry. However, the company run by Jensen Huang also competes in other sectors in which it is going more unnoticed.
One of them is the development of new lithographic technologies. In this area, NVIDIA develops Hopper architecture GPUs and cuLitho software libraries used by chipmakers to optimize their integration technologies. His role in this industry has led him to work closely with ASML and TSMC in fine-tuning the technological base that will make the production of 2nm chips possible. But there is something else in which this company is engaged and promises to have a lot to say: quantum computers.
With the error correction of quantum computers in the crosshairs
During the conference that he gave on March 21 to mark the opening of his event for developers, Jensen Huang talked about his collaboration with Quantum Machines. This Israeli company is specialized in the development of hardware and software for quantum computers, and has developed together with NVIDIA a low latency architecture and high performance that seeks to promote the advancement of quantum computing.
In this project NVIDIA has contributed its Grace Hopper CPU/GPU and its open source programming model CUDA Quantum
NVIDIA has provided its Grace Hopper CPU/GPU system, a beast that is designed to run AI applications and deliver productivity to match in high performance computing scenarios, as well as its open source CUDA Quantum programming model. Their partner in this project, Quantum Machines, has been in charge of integrating and fine-tuning a quantum platform that, according to these two companies, is specifically designed to work in hybrid systems in which classical and quantum hardware coexist in harmony.
The purpose of the DGX Quantum platform, which is the name of the hardware that these two companies have developed, is to help researchers working in the field of quantum computing to develop new quantum algorithms. It may seem surprising that it is possible to use classical hardware to develop quantum algorithms, but it is perfectly feasible. In fact, this strategy contributes to making quantum computing available to many more researchers who can implement and test their ideas without needing to have access to a quantum computer prototype, such as the one IBM is going to install in San Sebastián. (Spain).
However, the DGX Quantum platform is also used, according to NVIDIA, to calibrate quantum systems, control them, and even aspires to play a leading role in developing a correction system that allows quantum computers to correct their own errors. . Jensen Huang emphasized this idea during his keynote last week, and it’s certainly a very attractive prospect. Extraordinarily attractive. And, as Ignacio Cirac explained to us in the conversation we had with him, error correction will give us the opportunity to solve really significant problems with quantum computers.
More information: NVIDIA
In Xataka: Spain is going to have a 127-qubit quantum computer: it will be in the Basque Country and will be assembled by IBM
Leave a Reply