IBM optimizes quantum computing controls and accelerates workloads by 120 times
cbc b

Enhanced systems have reduced the amount of time per job run from 1,000 microseconds to 70 microseconds.

IBM announced on Tuesday (11) that its team demonstrated an acceleration of 120 times in the simulation of molecules while seeking an acceleration 100 times faster. The company links the feat to a number of improvements, including the ability to run quantum programs entirely in the cloud with Qiskit Runtime.

Blake Johnson and Ismael Faro, researchers at IBM, credited the improved performance to the improved control systems that reduced the amount of time per job execution from 1,000 microseconds to 70; and other performance improvements.

“Improvements to the algorithm have reduced the number of iterations needed to receive a final response by two to ten times. Improvements to the system software have been removed around 17 seconds per iteration. The improved performance of the processor led to a 10-fold decrease in the number of shots, or repeated circuit runs, required for each iteration of the algorithm, ”said the researchers and the company’s blog post. In addition to better reading and qubit reset performance, it reduced the amount of time per job run.

In 2017, the IBM Quantum team simulated the behavior of the lithium hydride molecule with a quantum computer, according to them, “a preview of the types of applications that we expect quantum computers to face in the future”. This modeling process took 45 days. Today, this process takes nine hours.

The researchers say the final boost came with the introduction of Qiskit Runtime, a containerized service for quantum computers. “Instead of accumulating latencies as the code passes between a user’s device and the cloud-based quantum computer, developers can run their program in the Qiskit Runtime execution environment, where the IBM hybrid cloud handles the work for them. New software architectures and OpenShift operators allow us to maximize the time spent on computing and minimize the waiting time, ”they said.

With that, the researchers hope that this acceleration will allow more developers to experience quantum applications in chemistry and other purposes.

“(…) Qiskit Runtime will allow users to try our new and powerful quantum kernel alignment algorithm, which looks for an ideal quantum kernel to perform machine learning tasks. We recently used this algorithm to prove that quantum computers demonstrate a rigorous acceleration over classic computers for supervised machine learning “, they point out.

The Qiskit Runtime is in beta for members of the IBM Quantum Network. Other milestones on IBM’s quantum roadmap include a 127-qubit IBM Quantum Eagle device, scheduled for this year, and the 1,121-qubit Condor device, planned for 2023.

Source: IBM

All comments.