30. January 2025
The second half of 2024 was an absolute blast as we worked with some of our close partners and customers to develop and release several new features to aid both modelling and bring-up of quantum systems. This release note highlights many of these features that are now out of beta and are considered stable and broadly available to all of our users. Most of these are in fact completely new and overhauled tools, so hopefully none of you are bored with bug-fix updates!
Contents:
The new version of the qruise-toolset
software library is carefully designed to simulate and optimise quantum systems in a way that allows for easy customisation and smooth calculations, without compromising on performance or scalability. Hamiltonian description is now straightforward, with direct support for QuTiP primitives for defining time-dependent and independent terms.
# define stationary Hamiltonian
H = Hamiltonian(sigmaz())
# define time-dependent Hamiltonian
H.add_term(sigmax(), drive)
A broad class of quantum dynamics problems – Schrödinger, von Neumann and Lindbladian – can be efficiently solved using our extensive library of ODE and piecewise constant solvers. Under the hood, it is powered by JAX which provides support for scaling large ensemble problems on CPUs & GPUs and also gives us automatic differentiation, necessary for fast optimal control. Read more about the new qruise-toolset here and check out the docs for some exciting new examples.
Before the second quantum revolution in computing and communications, NMR and MRI systems enabled powerful quantum sensing with broad medical applications. A significant part of the quantum optimal control toolbox we use at Qruise was originally developed by the NMR community in the early 2000s. A key area of ongoing research is the design of MRI scanners that provide the same level of accuracy and contrast for a fraction of the magnetic field (that's what's inside the large cylindrical enclosure in medical MRI scanners). QruiseML now supports modelling and optimising pulses for use in both traditional and hyper-polarised MRI applications. Stay tuned for updates in the coming weeks, where we'll showcase how our modelling library can be used to design optimal frequency-selective pulses for MRI.
Ion traps are one of the leading candidates for scalable fault-tolerant quantum computing and, over the last few months, we have added comprehensive support for modelling trapped ion systems. You can use qruise-toolset
for modelling standard Mølmer-Sørensen (MS)-type 2-qubit gates and optimising them with GRAPE for robustness to various experimental imperfections.
Additionally, we also support more advanced protocols such as the Light Shift gate proposed by G.J. Milburn in 2000, which is known to be less sensitive to some experimental imperfections. However, this protocol is susceptible to motional heating of the ion. Thankfully, the Qruise modelling tools allow you to include these phenomena and design pulses robust to their effect.
If you've been following us on social media (if not, do so now!), you'll have seen all the updates about the fantastic collaboration we have forged with our friendly colleagues at Quantum Machines (QM). Over the second half of 2024, we worked closely with the QM team and some early adopters to develop and test a robust and fully automated bring-up routine. We are now very happy to announce the general release of our integration with the QM control stack for all users. This enables quantum engineers to achieve high fidelity readout and single qubit gates from scratch in roughly 15 minutes, while accounting for factors like simultaneous sweet spots and coupler control in standard tunable-qubit-tunable-coupler superconducting architectures.
This integration with the QM control stack makes use of all the native optimisations (e.g., randomised benchmarking in a few seconds!) offered by the QM hardware & firmware and the qua python API. The extensive library of bring-up experiments combined with all the flexibility and scalability offered by the QruiseOS experiment management framework makes system bring-up an absolute breeze!
As QPUs scale to ever larger qubit numbers, it's imperative to perform bring-up experiments in parallel to be able to complete calibration and characterisation in a reasonable time. Otherwise, larger QPUs will experience significant downtime stemming from daily calibration runs. QruiseOS now supports fully parallelised and multiplexed calibration experiments, ensuring that bring-up times remain tractable when scaling to larger QPUs. We demonstrated this parallelised calibration on the Rigetti Novera device hosted at the Israeli Quantum Computing Centre (IQCC), which you can read more about here.
This support was first implemented at the individual bring-up experiments level, ensuring native parallelisation offered by control stack APIs is leveraged wherever possible. Additionally, the QruiseOS workflow definition file now also allows seamless switching between parallelised and sequential runs of bring-up experiments. Users now have the ability to scale parallel experiments to several qubits by simply listing the qubits in the workflow definition file.
name: demo-example-6-batch-tasks
qubits: [Q1, Q3]
batch_groups:
- name: all_qubits
qubits:
- [Q1, Q3]
stages:
---
experiments:
qubit:
---
- name: readout-discriminator-train
dependencies:
- time-t1
batch_group: all_qubits
We heard the feedback from our users regarding the need for a more streamlined GUI and we have now released a fully revamped dashboard. The homepage now packs a lot more information and key pages like System Info, Workflows, or the Experiment Database are just a single click away.
The workflow viewer has also undergone some improvements and there is a new sequential view (besides the original graph viewer) showing tasks in a waterfall as they progress. This new view also lets you filter the workflow by qubits, which becomes particularly useful when trying to isolate failing experiments in some naughty qubits!
Reach out today to request access if you'd like to try out all these new features in QruiseOS and QruiseML!
Subscribe to our newsletter to get the latest updates on our products and services.