IBM's quantum team cuts chemistry calculation from 45 days to 9 hours

IBM optimizes the quantum computing stack to make workloads 120 times faster by improving cloud service design and control hardware.


IBM's plan was to make quantum workloads 100 times faster at this point in the company's roadmap. Researchers expanded the scope of their improvement efforts beyond quantum circuits to the entire tech stack to exceed that goal. The company announced Tuesday that its quantum system has achieved a 120x speedup in simulating molecule activity. 

Blake Johnson and Ismael Faro described the improvements in a blog post on the company's site. The researchers credited improved control systems that reduced the amount of time per job execution from 1,000 microseconds to 70. 

In 2017, the IBM Quantum team simulates the behavior of the lithium hydride molecule with a quantum computer. This modeling process took 45 days. Today this process takes nine hours.  

SEE: Startup claims new "quantum analog computer" solved the traveling salesman problem for 128 cities (TechRepublic)

Johnson and Faro listed these improvements in the modeling process that allowed this significant acceleration:

  • Algorithmic improvements that reduced the number of iterations of the algorithm required to receive a final answer
  • Improvements in system software that cut about 17 seconds per iteration
  • Improved processor performance that reduced repeated circuit runs
  • Better readout and qubit reset performance reduced the amount of time per job execution  

Another element of the process that slowed down the workload was the back and forth between a classical algorithm and the quantum system, according to the blog post. The company credits Qiskit Runtime with reducing the latencies caused when code is passed from a user's device to the quantum computer in the cloud. 

Qiskit Runtime completes the classical and the quantum elements of the analysis in the cloud. Johnson and Faro wrote in the blog post that this new feature should open up quantum computing to more developers:

".... the Qiskit Runtime will allow users to try out our powerful new quantum kernel alignment algorithm, which searches for an optimal quantum kernel with which to perform machine learning tasks. We recently used this algorithm to prove that quantum computers will demonstrate a rigorous speedup over classical computers for supervised machine learning." 

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

Quantum computing is also a topic at IBM's THINK 2021 conference. Leaders from Anthem and BP will present their quantum work at the conference which is May 11 and 12. IBM will share an update on its quantum roadmap as well as its work with historically Black colleges and universities to develop a quantum ecosystem.

Also see