[GH-ISSUE #25] Chome Flag investigations for speed and success (Aditional chrome startup flags for complex pages or low power Pi Zero W) #22

New issue

Closed

opened 2026-03-03 11:13:26 +03:00 by kerem · 4 comments

kerem commented

2026-03-03 11:13:26 +03:00

Owner

Originally created by @andiohn on GitHub (Jul 30, 2024).
Original GitHub issue: https://github.com/debloper/xiosk/issues/25

Taken directly from this issue, could be huge?

https://github.com/jareware/chilipie-kiosk/issues/197

Originally created by @andiohn on GitHub (Jul 30, 2024). Original GitHub issue: https://github.com/debloper/xiosk/issues/25 Taken directly from this issue, could be huge? https://github.com/jareware/chilipie-kiosk/issues/197

kerem

2026-03-03 11:13:26 +03:00

closed this issue
added the
good first issue
label

kerem commented

2026-03-03 11:13:27 +03:00

Author

Owner

@andiohn commented on GitHub (Jul 30, 2024):

Other flags to investigate:
https://github.com/jareware/chilipie-kiosk/issues/181

@andiohn commented on GitHub (Jul 30, 2024): Other flags to investigate: https://github.com/jareware/chilipie-kiosk/issues/181

kerem commented

2026-03-03 11:13:27 +03:00

Author

Owner

@debloper commented on GitHub (Aug 1, 2024):

Added these following flags (patch landing soon):

  --fast \
  --fast-start \
  --enable-accelerated-video-decode \
  --enable-gpu-rasterization \
  --enable-oop-rasterization \
  --ignore-gpu-blocklist \
  --disable-smooth-scrolling \
  --disable-low-res-tiling \
  --enable-low-end-device-mode \
  --disable-composited-antialiasing \

It's hard to do thorough differential benchmarking to validate each flag's impact on performance, and it's even harder to choose the specific pages as sample to benchmark on. For example, time.is is super easy on resources (1-2% CPU usage on RPi4), whereas windy.com takes up over 50% of CPU on all 4 cores while active.

And even with that much of disparity I didn't see much of a difference on resource footprint (with htop) between the before and after the flags being applied. Also, some of the flags may even be deprecated, for example: --disable-features=OverscrollHistoryNavigation.

However, there's nothing misbehaving, so I'm leaving them in. Might be better to split them up (required vs optional) as part of config.json & configurable by the user through the dashboard.

@debloper commented on GitHub (Aug 1, 2024): Added these following flags (patch landing soon): ```bash --fast \ --fast-start \ --enable-accelerated-video-decode \ --enable-gpu-rasterization \ --enable-oop-rasterization \ --ignore-gpu-blocklist \ --disable-smooth-scrolling \ --disable-low-res-tiling \ --enable-low-end-device-mode \ --disable-composited-antialiasing \ ``` It's hard to do thorough differential benchmarking to validate each flag's impact on performance, and it's even harder to choose the specific pages as sample to benchmark on. For example, `time.is` is super easy on resources (1-2% CPU usage on RPi4), whereas `windy.com` takes up over 50% of CPU on all 4 cores while active. And even with that much of disparity I didn't see much of a difference on resource footprint (with `htop`) between the before and after the flags being applied. Also, some of the flags may even be deprecated, for example: `--disable-features=OverscrollHistoryNavigation`. However, there's nothing misbehaving, so I'm leaving them in. Might be better to split them up (required vs optional) as part of `config.json` & configurable by the user through the dashboard.

kerem commented

2026-03-03 11:13:27 +03:00

Author

Owner

@andiohn commented on GitHub (Sep 30, 2024):

https://pimylifeup.com/raspberry-pi-kiosk/

This one has a lot more options too, some of them look good, like avoiding the errors etc.

@andiohn commented on GitHub (Sep 30, 2024): https://pimylifeup.com/raspberry-pi-kiosk/ This one has a lot more options too, some of them look good, like avoiding the errors etc.

kerem commented

2026-03-03 11:13:27 +03:00

Author

Owner

@debloper commented on GitHub (Oct 7, 2024):

Error suppression wouldn't be enabled by default, but user config to enable/disable certain startup flags can be implemented.

@debloper commented on GitHub (Oct 7, 2024): Error suppression wouldn't be enabled by default, but user config to enable/disable certain startup flags can be implemented.