Researchers are overcoming challenges in quantum computing by integrating artificial intelligence (AI) methods with quantum processing units (QPUs). This accelerated quantum supercomputing approach enables breakthroughs in fields like compact circuit synthesis and quantum error correction. However, accessing diverse QPU hardware from various vendors has been a persistent problem due to differing access procedures, software stacks, and pricing models.
To address this issue, NVIDIA is collaborating with Amazon Web Services (AWS) to expand access to quantum hardware by integrating NVIDIA’s CUDA-Q platform with AWS’s Amazon Braket. This collaboration allows researchers to tap into powerful GPU computing resources and a wide array of QPUs without needing contracts with specific vendors. Key individuals involved in this work include those from NVIDIA and AWS, working together to advance hybrid quantum computing research.
This integration enables researchers to experience the best of both worlds by leveraging the performance of CUDA-Q and the flexibility provided by Amazon Braket. With this collaboration, researchers can now access a range of QPUs, including those from IQM, without the need for vendor-specific contracts.
The code snippets demonstrate how easy it is to prepare CUDA-Q code within the Amazon Braket experience. By setting the target to “braket” and specifying the machine parameter as the device, researchers can run their CUDA-Q kernels on various QPUs available on Amazon Braket. This flexibility allows for faster experiment times and minimal hardware drift, improving results.
One key benefit of this integration is that no upfront fees or commitments are required to use QPUs on Amazon Braket. This means researchers can get started using QPUs without entering into contracts with individual hardware vendors.
The article highlights two examples of accessing CUDA-Q within the Amazon Braket experience. The first example shows how to run a Bell state CUDA-Q kernel on a specific QPU, such as IQM‘s Garnet superconducting QPU. The second example demonstrates how to create a hybrid job that takes a CUDA-Q simulator backend as the device, along with an instance that specifies the GPUs to be accessed for the computation.
The procedure for accessing one of Braket’s available QPU backends is similar, involving defining a target QPU from Braket and specifying it as the device for the hybrid job. The only other change required is to set the CUDA-Q target to “braket” along with the machine parameter, which is set as the device.
This integration has significant implications for advancing hybrid quantum computing research. By combining the performance of CUDA-Q with the flexibility provided by Amazon Braket, researchers can now experience the best of both worlds. I’m excited to see how this collaboration will accelerate progress in the field of quantum computing.
External Link: Click Here For More
