New Product Releases: Pasqal's Pulser Studio, IBM's Sherbrooke Processor, and Google Cirq v1.1.0

Several interesting new product releases were made over the past few weeks including Pasqal's Pulser Studio no-code development platform, IBM's 127 qubit Sherbrooke processor, and a new release of Cirq version 1.1.0.

Pasqal's Pulser is a platform that allows users to create registers and pulse sequences for a neutral atom quantum computing using a graphical interface and not require any coding knowledge. The platform allows a user to put together a quantum program based upon five element: Atoms, Registers, Pulses, Channels and Measurements. It is open and free to corporate and academic users who can sign up for the beta release and includes a built-in emulator for simulating small systems. The goal of this platform is to make it easier and faster for end users to program their processors while helping them develop insights into how the program will work. Later this year, Pasqal will integrate this software into their cloud computing platform so it can be used with real quantum processors as well as making updates to add additional features. More information about Pulser Studio is available in a new release located here, a web page available here and a beta user sign-up page here.

Pulser Studio Screen Shot. Click on the picture to view the video. Credit: Pasqal

Right before the holidays, IBM put on line their 127 qubit IBM_Sherbrooke processor which they say is their highest performance system to date. Sherbrooke is part of their Eagle processor family introduced last year and is a third iteration known as Eagle r3. Their first 127 qubit processor is known as IBM_Washington and is denoted as Eagle r1. Although it has the same number of qubits, Sherbrooke has been optimized for error mitigation for techniques such as Probabilistic Error Cancellation (PEC) and Zero Noise Extrapolation (ZNE) to help provide better quality programming results. The first notable difference we see, as shown on our Qubit Quality page, is an improvement in T1 coherence times from about 95 microseconds with Washington to 312 microseconds with Sherbrooke. The T2 times also improved from 97 microseconds to 177 microseconds. Additional improvement can also be seen in the single gate fidelity, 2-qubit gate fidelity, readout, and CLOPS measures.

IBM has also changed their calibration strategy with this device. Previously they designed their calibration routines to minimize the error rates, but at the expense of stability. With this device they are changing that to emphasize reducing measurement leakage, gate stability and more uniform gate speed. This will help eliminate drift and extend the time before another recalibration is required. Interestingly, this new calibration strategy may not necessarily provide the best possible quantum volume measurement but may still be preferable because of the increase in stability. Along with this change, IBM has implemented a different 2-qubit native gate called the echoed cross-resonance (ECR) gate to replace the CNOT 2-qubit gate previously used. The Qiskit transpiler has been updated to compile any existing Qiskit program to use this gate, but it also makes this gate directly available in Qiskit for programmers who want to use it. The topology of Sherbrooke is shown below. Note that the 2-qubit ECR gates is only unidirectional (as shown below by the connection arrows) rather than bidirectional. The transpiler knows this and will adjust for this when it allocates the physical qubits while compiling the program. IBM has posted a new blog about this processor that you can find here and the updated comparison listing on our Qubit Quality page here.

Topology of the IBM_Sherbrooke (Eagle r3) Processor. Credit: IBM

Finally, Google has released a Cirq v1.1.0 of their open source quantum programming framework. This Cirq release focuses on tracking and improving performance of key workflows like circuit construction, parameter resolution etc. The release also adds a new transformers framework for qubit routing and provides an efficient implementation of a qubit routing algorithm developed originally by researchers at Cambridge Quantum Computing (now Quantinuum) and others which is described in arXiv:1902.08091. Additional information about the changes in this release is available on the Cirq v1.1.0 GitHub page here.

January 7, 2023