NVIDIA is pushing the idea of combining elements of quantum and classical computers to speed up computation.
In particular, the company is working to incorporate its graphics processing unit (GPU) computing gas pedals into quantum systems, as well as to simplify the creation of quantum algorithms.
And today NVIDIA has announced new steps on that path.NVIDIA some time ago introduced the cuQuantum toolkit, which allows the simulation of algorithms for quantum processors on NVIDIA A100 GPUs, and more specifically on their tensor cores.
This made it easier to enter the world of quantum computing.
QuQuantum, in particular, is available to everyone as part of the AWS cloud.
Now NVIDIA intends to merge quantum and classical systems.
To do this, NVIDIA wants to create a low-latency interface that will allow it to link its quantum computing gas pedals and quantum processors (QPUs).
This would allow quantum computers to use the powerful parallel computing potential of GPUs to solve classical problems.
In particular, it is proposed to use them for circuit optimization, calibration and error correction.
GPUs can reduce the time it takes to perform these tasks and reduce communication latency between classical and quantum computers, a major bottleneck for today's hybrid quantum systems.NVIDIA also believes the quantum computing industry needs a unified programming model with tools that are efficient and easy to use.
Today, researchers are forced to use the quantum equivalent of low-level assembly code to program QPUs, which is beyond the capabilities of many scientists.
NVIDIA will offer a suite of tools that allow scientists to easily implement their quantum algorithms, first on simulated QPUs, then on real QPUs.
What's needed is a compiler that can run in both environments.
By combining GPU quantum computing simulation tools with a unified programming model and compiler, researchers will be able to begin building hybrid quantum data centers, according to NVIDIA.
0 Comments