Because evolution is unitary.
Results Are in for the 2022 Quantum Open Source Software Survey!
Unitary Fund Team
We are excited to share the results for the 2022 Quantum Open Source Software (OSS) Survey organized by Unitary Fund. Collectively, over 1000 responses were provided by the community to the main software survey (Part I) and to the community survey on Diversity & Inclusion (Part II).
You can view the results in full screen mode at this link.
Part I of the survey was divided into the following sections:
- Experience (including: Cloud services; Full-stack development platforms and simulators; Software for applications and tools)
- OSS Development & Research
Let’s highlight some of the most interesting results about quantum open source software users and developers.
Demographics: While the majority of quantum OSS users are researchers (55%), sizable communities identify themselves as software developers (34%), students (31%), educators (15%) or hobbyists (15%). This data speaks for the balanced heterogeneity of interests and sub-communities among quantum OSS users and developers. Almost 40% of respondents do not have a background in quantum research.
While about half of respondents are affiliated to an academic institution (49%), about one third work in an enterprise organization (29%) and a quarter in a startup (25%), with 9% having no affiliation and 8% working in government.
The most represented country is the United States (22%), with India coming in second (12%), the United Kingdom at 9%, and Canada at 6%. EU countries sum up roughly 15%, with Germany first among them at 5%. This data certainly speaks for the enthusiasm that quantum computing is generating in India where, compared to other major countries, fewer institutional or tech transfer initiatives have been developed (e.g, fewer startups or regional programs compared to Europe).
The majority of respondents either work full time (53%) or part time (10%) in the quantum industry, and among those, about 20% work fully remote, about 40% employ a hybrid format, and only 16% are fully in-person, possibly a signal of recent work format changes enabled by technology and accelerated by the pandemic.
Experience: About 90% of respondents use quantum software, of which about half are solely users, and 45% are either OSS project contributors, maintainers, or owners. Turning to Cloud services connecting to real quantum processing units (QPUs) or simulators, about 80% of respondents have used them in the past year: That’s definitely a high percentage. The most popular service is clearly IBM Quantum (80% of respondents are current users), most likely leveraging their position as the first major provider of QPUs and providing free access to many devices. IBM is followed by AWS Braket (21% are current users, and 27% would be interested in trying it out in the next 12 months), with Xanadu at 16%, Microsoft Azure Quantum at 15%, and Google at 15%. Users with Rigetti Cloud Services (7%), IonQ (8%) and Honeywell Quantum (8%) are smaller through their individual cloud services, but one should keep in mind that their QPUs are popular and available also through the major cloud service providers. It will be interesting to see if in the future “vendor-lock-in” effects will consolidate the cloud offer or a multiplicity of cloud-access providers continue to build their own platforms will keep broad. Maintenance, documentation and price are major factors for users to decide whether to use a cloud service or another one.
With regards to Full-stack development platforms, respondents have indicated that IBM’s Qiskit (including Qiskit Aer) is their most popular library (81%), followed by Google’s Cirq and Xanadu’s PennyLane (29%) and then by AWS SDK in Python (16%) and tket (15%) (Quantinuum). There is interest in AWS SDK for the next 12 months. Among libraries with more than 10% of users there are Strawberry Fields and QuTiP-QIP (an affiliated project of Unitary Fund and the only project of these not directly backed by a startup or corporate). Other popular libraries are Dwave’s Ocean SDK, Q# (Microsoft) and cuQuantum (Nvidia), Rigetti’s pyQuil and Quirk. Documentation is deemed very important or important by about 90% of respondents as a factor to weigh in when choosing an SDK.
With regards to tools for applications, Qiskit packages such as qiskit-optimization, qiskit-machine-learning, and qiskit-nature are among the most popular ones, followed by PennyLane’s QML repo, OpenQASM, qiskit-finance, tensorflow-quantum and Unitary Fund’s Mitiq for quantum error mitigation. Other popular projects include OpenFermion, the PyZX compiler, stim for quantum error correction. There is widespread interest for using other tools in the future such as covalent, qcor, NetKet, ScaffCC, orqviz, tweedledum, QuNetSim, SimulaQron, qmasm, and toqito: these projects range from compilers to network simulators to specific tooling.
With regards to OSS development and research, over 40% of respondents performing research define themselves as algorithm or applications developers, with about one third involved in circuit development & optimization or quantum simulation/Physics. A sizable percentage are involved in software development (30%) and quantum information theory (26%). Other interests include fundamental physics and quantum error mitigation, while less than 13% declare they are performing research in either quantum error correction, qubit characterization, or hardware: that’s definitely surprising especially given the importance of QEC and software integration with hardware for the advancement of the field.
The most popular programming language is Python, and this was to be expected, but it is quite impressive that over 94% of respondents use it, while the second most popular framework is far behind, C/C++ (26%), with Mathematica and MATLAB tied at 10%, together with the Julia language. Rust appears at 7%. It will be interesting to see how a scientific language like Julia and a performant language like Rust will change in adoption next year. Jupyter Notebooks and notebooks in general are very popular as tools for software development (used by 78% of respondents), with 68% of respondents using an integrated development environment (IDE) and 52% using the command line or terminal.
With respect to the quantum software community, 56% find it very welcoming, 28% somewhat welcoming and 10% neither welcoming nor unwelcoming or worse. 94% of respondents have a positive view of OSS in the quantum software community, with 78% finding it has a very positive impact and 16% a somewhat positive impact. With respect to the most sought after sources of answers or information when developing quantum software, project documentation or project websites are the two most popular (84%), with project repositories and codebase coming in third. Standard forums are slightly less popular nowadays (8%), when compared to platforms such as the Quantum Computing Stack Exchange, Stack Overflow, and servers such as Slack and Discord, and Youtube videos. It is also quite surprising to read that among the types of resources most helpful for learning or contributing to quantum open source projects, video resources fare first (67% find them useful), followed by digital education text resources, hackathons, participative courses, mentorship programs and certificate or degrees (all above 40%).
Open-ended Feedback on the quantum OSS ecosystem and community from respondents can be found at this link. We highlight just a few below, that emphasize the work needed and requested to support quantum open source software:
“I wish that everyone could assist [the] quantum OSS community to accelerate its development.”
“I think there are projects of high quality that are unfortunately underrepresented because they lack polish, industrial support, and/or advertising.”
“I’d love to see more community-building; interviews with the personalities behind the tech that emphasize the humanity over the math and physics. We already have the math and physics bits, but do we really know each other?”
“I would like to see a quantum sdk that is not owned by a company and is of high quality. All research-related projects I have seen were not easy to use.”
“In my experience the supportiveness and inclusivity of the quantum OSS community has been wonderful (I have most experience with Unitary Fund and Qiskit communities). As a student, finding a good and supportive teacher/mentor has always been the thing that decides which subject areas I end up falling in love with. The same has been true for learning about quantum computing, and quantum OSS. [..] the extent to which the quantum OSS community is welcoming and encouraging (and correspondingly does not fetishize "brilliance", "innate intellectual ability", or formal training in the way many STEM fields do)–will determine, more than anything else, the extent to which quantum computing can become a diverse, open and truly meritocratic (in the positive sense) field in the future. The idea that this kind of social-political progress in STEM can be part of what QC contributes to humanity is part of what motivates me to pursue a career in QC (in addition to the intrinsic fascination/wonder of QC, which is for me the primary reason).”
“It’s something I’ve noticed in traditional open source software, and something I hope doesn’t carry through to quantum open source software: OSS is taken for granted and doesn’t get the recognition it deserves (which has hindered it), I think that we should publically recognise and celebrate the impact of Q-OSS, even if it’s under a proprietary company (eg Cirq, Qiskit etc)”
Community: Diversity & Inclusion
Part II of the survey (shared as a separate survey for anonymity and hence in principle with a different or overlapping pool of respondents) focuses on providing a snapshot of the diversity and inclusion of the quantum software community. Similarly to the software survey of Part I, it shows that the USA and India are the most represented countries of residence. About a half of respondents are in the 25-34 age range, with 20% of respondents below 25 (but only 1% under 18 years old).
With regards to ethnicity, about a half of respondents identifies as White or of European descent, 12% as South Asian, 8% as Hispanic or Latino/a/x, 6% as Black or of African descent, and 4% or below as East Asian, South East Asian, Multiracial, Middle Eastern, Other, or prefers not to say. 73% of the respondents identify as male, 19% as female, 3% as non-binary, genderqueer, or gender non-conforming, 1% other and 7% prefer not to say. With respect to educational background, the largest group holds a PhD (36%) or multiple ones (2%), 33% have Masters or other non-doctoral post-graduate degree, 22% have a university degree and 8% high school/secondary school degree, other degrees, or prefer not to say. With respect to sexual orientation, 77% are straight/heterosexual, 9% bisexual, 9% prefer not to say, 2% are queer, 2% are gay or lesbian, and 1% have another orientation.
With respect to physical disabilities, 86% do not report any, while 7% prefer not to say and 5% replied other. 1% of respondents are blind or have difficulty seeing, are deaf or hard of hearing or unable to(/find it difficult to) walk or stand without assistance. With regards to mental health, 66% are not diagnosed with any mental health condition and 12% preferred not to say. Other respondents selected anxiety (12%), mood or emotional disorders (9%), concentration and/or memory disorder (8%), other (2%), autism (1%), or schizophrenia (1%).
The survey has been open Sept. 7 – Oct. 7, 2022. The data is stored at github.com/unitaryfund/qoss-survey.
Unitary Fund circulated the surveys on its social media platforms (Discord, Twitter, LinkedIn, UF blog) and contacting major blogs (QC Report, Qiskit blog, PennyLane blog, etc.), newsletters (UF mailing list, QuTiP mailing list, ORNL quantum computing newsletter, IEEE Quantum, QED-c newsletter, academic networks, etc.), UF members, supporters and partners.
We are excited to repeat the survey in coming years and track the changes and trends in responses and in the field. Thank you to all that have participated!