Norton interfering with pandas and numpy load (python development)

Hi all - hope your days are going well :smiley:

Issue abstract: A recent update to Norton is slowing the loading of pandas and numpy packages used by my long-established python apps.

Detailed description:

I develop applications in python, and the packages pandas and numpy are commonly involved. Normally, these packages load in less than a second (including all their dependencies) but recently I have noticed that after the machine has been started for the first time in a day, or after a restart, the first load is very long (minutes). If I rerun any of the apps, then the load is instant.

I have spent several hours narrowing down what is going on, and it is clear that it is Norton real-time protection that is involved. Below are the timings I have recorded using Terminal, activating the .venv environment for one of my projects, then running bash commands to capture timings. So cold means this is first run after a restart. Warm means a second run. As you can see, There is a big difference turning Norton protection off.

Python startup ~0.03 s
NumPy cold, Norton ON 21.7 s
NumPy warm 0.09 s

pandas cold, Norton ON 178.19 s
pandas warm 0.28 s

pandas cold, Norton Auto-Protect OFF 0.775 s
pandas cold, Norton ON + venv excluded 0.722 s

My temporary fix is to exclude the .venv directories for apps that use these packages, but that isn’t ideal.

To me, it appears that the behaviour of Norton real-time protection ie. its Endpoint Security component, its signatures, or its caching policy etc. As I mentioned, these are long-standing apps with known and predictable behaviour.

Anyone else seeing this? Any solutions out there?

Product & version number: Norton 360 version 26.8.1

OS details: MacOS 26.6.2 (25G83). Running on MacBook Air M3.

What is the error message you are seeing? No error message

If you have any supporting screenshots, please add them:

While Norton might have contributed to a slight delay in loading panda and numpy packages because of it’s real-time protection, I don’t see such a big difference. The difference is almost none. Maybe, that’s because of my higher spec. machine? My spec: M2 Max 16-inch MBP with 32 GB memory, macOS Golden Gate

Thank you for the reply on this particular issue. I’ve done some more investigation and come to the conclusion that my Norton installation is corrupted. I’ve reinstalled and the situation is vastly improved. However, I have discovered that the latest pandas and numpy libraries running under Python 3.14 are markedly slower to load then some of the older libraries. I don’t think I can blame Norton on that one! Again, many thanks for the reply and good luck with Golden Gate :grinning_face:

1 Like