The announcement was made during Jensen Huang’s introduction to the virtual Graphics Technology Conference earlier this month. Huang is the co-founder and CEO of Nvidia.
Huang described cuQauntum as an acceleration library designed for the simulation of quantum circuits that is optimized to scale to large GPU memories and multiple GPUs and DGX nodes. DGX is a server that comes with Intel CPUs and Nvidia GPUs and operating system software.
Developers can use cuQuantum to speed up quantum circuit simulations based on state vector, density matrix, and tensor network methods.
Using the state vector method a DGX-A100 with eight A100 GPUs can simulate up to 36 qubits and reduce processing time from 10 days to a couple of hours.
The tensor network method trades memory footprint for run time allowing larger quantum circuits to be simulated but at reduced fidelity to the state vector approach.
In his introductory address Huang said: “I’m hoping cuQuantum will do for quantum computing what cuDNN did for deep learning.”
Related links and articles: