Quantum computers don’t all work the same way, and that creates a problem for developers. Code written for one machine may need to be rewritten before it can run on another. Postquant Labs wants to change that.
The US startup has launched the Quip Virtual Machine (QVM), a software layer that lets developers write code once and run it across different types of quantum computers and conventional machines.
The QVM is available on the company’s Aglais testnet, where new users receive free computing credits. It runs on Quip Network, a quantum computing marketplace that Postquant Labs opened for testing in April.
At the moment, major quantum computing providers have their own programming tools and pricing systems. Developers need to learn how each one works and adapt their code accordingly. Postquant Labs wants to simplify that process by providing a common way to work across different machines.
Programs written for the QVM can run on quantum annealers, which are designed to find good solutions to optimisation problems, gate-based quantum computers, and conventional computers.
“Quantum computing is still fractured between competing instruction sets and assemblers, providing few easy onramps to outsiders,” said Colton Dillion, CEO and co-founder of Postquant Labs.
The company is also trying to make computing costs easier to predict. Its pricing system borrows from Ethereum, using a unit called gas to calculate the cost of each job.
The price depends on factors such as the number of qubits used, the time a job takes, and the amount of entanglement it requires. Developers can see the maximum cost before running a job, the company said.
Postquant Labs has also built in ways to check computing results. These include comparing quantum outputs with simulations on conventional computers, having other providers verify results, and digitally signing each result to identify the machine that produced it.
At the Quantum.Tech World conference, the company tested the QVM on an investment portfolio selection task, running 1,930 comparisons between quantum and classical solvers.
Postquant Labs said its quantum solver matched or outperformed the classical one in every comparison. When the performance gap exceeded 0.5%, the median advantage was 2.56%, according to the company.
The results are company-reported and do not establish that quantum computers generally outperform conventional ones on investment tasks.
Researchers at Monash University, the University of Notre Dame, and Purdue University are testing the QVM, according to Postquant Labs. SiC Systems, an industrial AI company, is also exploring its use for factory planning.
The platform will let developers publish programs and earn fees whenever others run them. Once Quip Network is fully operational, users will pay for computing jobs with network tokens.
Postquant Labs is also launching QuipSwap, a payment system designed to withstand attacks from future quantum computers, to handle transactions.