Since Borealis is inaccessible and X8 device can only support a few specific gates and states. I wanted to implement CVQNN on a real device or any circuit for that matter, to see how good the continuous variable-based computing is. Now, it is very difficult to simulate the circuits using strawberry fields due to the large dimensions they take, and has now become impossible.
Is there a universal photonic quantum computer yet, because X8 also seems to be performing specific tasks? For example, there are no Kerr Gate or Displacement Gates that are essential for doing CV computation. I would also love to know if any alternatives are there.
Do we expect in the near future for these problems to be solved so that I can start sending jobs through strawberry fields?
Indeed the X8 device is from 2019 so it looks small now in 2023. It’s mostly useful if you want to do some specific research in photonics.
For CVQNN at the moment the option that will give you highest performance is using a simulator (which you can find in Strawberry Fields). As expected you will reach limitations if you want to simulate large and complex systems.
Specifically responding your questions:
- We are indeed working on building a universal photonic quantum computer. However …
- Such computer is based on a different architecture. It uses measurement-based quantum computing so it won’t work like the X8. It will still be a photonic quantum computer but not CVQC.
You can learn more about measurement-based quantum computing in this demo.
Please let me know if you have any further questions!
Indeed the X8 device is from 2019 so it looks small now in 2023
I understand that it is smaller than the current standards, but still having 8 qumodes is still a big deal, as you can’t easily simulate those 8 qumodes in a classical machine. However, it only supports a limited number of Gaussian gates and that is something I wanted to confirm.
It will still be a photonic quantum computer but not CVQC.
So can I expect a qubit-based universal quantum computer to be made soon? (from the different academic talks, it is computation using GKP states, that I heard being developed.) I understand if this is something confidential.
In conclusion, all I can do is wait, given that I can’t simulate circuits any further, isn’t it?
Thank you for the time and insights
Hi @Prabhat_Prabhat ,
Yes, X8 is still useful for education and research but as you saw the gates available are limited.
Regarding your second question, yes, the device we’re working on is qubit-based and universal! And it does indeed use GKP states.
If you’re interested in simulations there’s still a lot of ground to explore. You can explore our different libraries (Mr Mustard, The Walrus, FlamingPy, Strawberry Fields, PennyLane) and see how you can leverage them to do the simulations you want. I suggest you first go through the documentation to understand what they do and then if you have any specific questions feel free to post a question here in the Forum.
Let me know what you think of these libraries!