We were looking into performance improvements for registering notebook controllers for Python kernels and while debugging we found that

• https://github.com/microsoft/vscode-python/blob/main/src/client/interpreter/interpreterService.ts#L139 InterpreterService#getInterpreters is invoked four times (on my machine particularly) and from the log, each time it returns the same three interpreters and each call takes one second.

Under the hood it's using fs to read and creating processes to find intepreters so maybe there is room for perf improvement to do some caching or moving some of the code to external module/binary.

Environment data

• VS Code version: 1.59-insiders
• Extension version (available under the Extensions sidebar): 2021.7.3-pre.1
• OS and version: Windows 10
• Python version (& distribution if applicable, e.g. Anaconda): Conda, Pyenv
• Type of virtual environment used (N/A | venv | virtualenv | conda | ...): pyenv, conda