Date: Saturday, March 1, 2025
Time: 01:20 PM - 5.40 PM Pacific Time (PDT) — UTC-8
Room:
Time: 01:20 PM - 5.40 PM Pacific Time (PDT) — UTC-8
Room:
The last few years have seen quantum computing workflows starting to enter HPC environments. On the one hand, GPU-accelerated quantum simulations are increasingly being adopted in hybrid quantum-classical algorithm development to speed up algorithm run time and to scale up the size of quantum research. On the other hand, new workflows are emerging that require close coupling between quantum processing units (QPUs) and GPUs.
This interactive tutorial guides attendees through new tools and developments in the field, focusing on the CUDA-Q software environment and hybrid workflows with both simulated and real QPUs. The tutorial begins with an introduction to CUDA-Q and an interactive Jupyter notebook demonstrating parallel quantum simulation using NVIDIA Quantum Cloud. Next, attendees will learn about photonic quantum computing, how this modality is supported by CUDA-Q, and about ORCA Computing's photonic quantum processors. Finally, the Poznan Supercomputing and Networking Center (PSNC) will present their hybrid multi-GPU and multi-QPU HPC environment, and enable attendees to deploy quantum software on a large-scale HPC cluster involving two ORCA Computing QPUs and multiple GPUs.
You can attend most of this tutorial without registering ahead of time, however you do need to register ahead of time to participate in the hands-on part. This is required for access to PSNC's compute resources. To register, click on the registration link above before the 21st of February.
Outline of the program
Introduction to CUDA-Q: build and simulate quantum kernels on CPUs and GPUs, emphasize differentiating features from other SDKs, recognize and employ various parallelization strategies across various target backends and implement them on NVIDIA Quantum Cloud and PSNC architecture. We will provide examples of how to use CUDA-Q to solve problems in optimization using QAOA. The topic of using generative AI to generate optimal ansatzes will be covered.
Quantum Hardware: Introduction to photonic quantum computing and to ORCA Computing's photonic QPUs. We will present how to program photonic quantum processors using CUDA-Q, discuss some applications, and show how these can be deployed within a multi-QPU, multi-GPU workflow with 2 ORCA PT-1 systems recently installed at PSNC in conjunction with NVIDIA hardware.
Intro to large-scale clusters at PSNC: Introduction to PSNC's hybrid quantum/classical HPC environment and the algorithms that were run in this environment.
Hands-on session: introduce the audience to logging in, running interactive sessions, submitting jobs, and post-processing results on supercomputers with PSNC as the point of reference. This will include showing users how to deploy docker and use Jupyter notebooks with hands-on access to PSNC. We will also show the audience how to scale simulations from 1 to 100s of GPUs with minimal changes in workflow. An example of a Hamiltonian with 30,000 terms will be demonstrated and how this can be easily parallelized using CUDA-Q. The audience will have a chance to follow along by having access to the infrastructure at PSNC.