Document NumPy 1 interaction with max simulation size #1015
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request adds a helpful note to the documentation regarding issues with NumPy 1.26 in large simulations, recommending an upgrade to version 2.x. However, this creates a conflict with the requirements.txt file, which still permits the installation of NumPy 1.26. My review includes a comment to address this inconsistency to prevent potential issues for users.
| * NumPy version matters for large simulations: | ||
| * Some users have reported being unable to simulate circuits with more than | ||
| ~32 qubits when using NumPy 1.26. | ||
| * Updating NumPy to version 2.x resolved the issue in these cases. | ||
| * If you encounter unexpected failures, memory errors, or hard limits on | ||
| the number of qubits, ensure you are running a recent NumPy version | ||
| before further debugging. |
There was a problem hiding this comment.
This documentation change introduces an inconsistency with the project's dependencies. While it correctly warns about issues with NumPy 1.26 for large simulations, requirements.txt still specifies numpy>=1.26.0. This can lead to users installing a version known to have problems for certain use cases, creating a frustrating experience.
To maintain consistency within the project, please update requirements.txt to align with this new recommendation. For example, you could change the dependency to numpy>=2.0. If this is not feasible, the documentation should explain the trade-offs.
References
- The repository style guide emphasizes consistency across the codebase. The documented NumPy recommendation conflicts with the version specified in
requirements.txt, creating an inconsistency. (link)
1 participant
#863