Regression in 4.0.0 -> 4.1.0, Live graph performance on Linux

[Manawyrm]

Hi,

side note: I've signed up for DevZone, the provided account credentials did not work work login after account creation. No idea why, that's why you're getting the report here. Feel free to move it over to DevZone and close this issue if preferred.

Updating the Power Profiler software from v4.0.0 to v4.1.0 caused the Live view/graph performance to become much, much worse.

Starting with v4.1.0 (and also in the current 4.2.0), the graph is only updated in "chunks":
https://github.com/user-attachments/assets/bcf4b85d-20b9-48bf-900c-b65ff3509ef2

The expected behavior (with v4.0.0) was:
https://github.com/user-attachments/assets/885a2db4-6dd4-407a-ad6a-5baf3ddf2c1e

Additional info

• Manjaro Linux (Arch Linux-based)
• nRF Connect installed via https://aur.archlinux.org/packages/nrfconnect-appimage
• Linux kernel 6.10.5

The pull request #449 by @kylebonnici might be related here (but I have no way to test this).

Thanks for making this excellent tool (both in application & hardware)!

[Manawyrm]

Using the device further, with both v4.0.0 and v4.2.0 there seems to be something very, very wrong with the timebase/time measurement.

2024-08-24.21-53-13.mp4

10s of wall clock time is measured as 3 seconds in PPK. Curious!

[Manawyrm]

Ah, #318 is the root cause here.
Putting the PPK2 behind a USB hub (instead of the mainboard ports directly) does indeed fix the issue.

Unfortunate bug...

[Manawyrm]

This also fixes v4.1.0/4.2.0's display behaviour!

Closing this issue, adding my feedback to #318.