Inquiry regarding the performance improvements done from version 1.87 release of Foxglove.

[shub14]

I've recently started using the Lichtblick for ADAS-related visualisations, and I am facing a lot of lag while using the large .mcap files. I have done some research and found that Lichtblick has been forked from Foxglove. I checked and found that Lichtblick has been diverged from v1.86 release of foxglove. I have seen that foxglove had done some performance improvements in v1.87 release. Are these changes have been added within Lichtblick as well or have you guys did any other optimizations?

I am attaching screenshot and link for v1.87 release for your reference?

Link - https://github.com/bgromov/foxglove-studio

[aneuwald-ctw]

Hi @shub14, thanks for reaching out.

Some of them were integrated, but the buffering messages as UInt8Array we haven't implemented. I give a quick try and indeed it increase a lot the performance! At least in the use cases that de buffering was very heavy.

We will proper bring this to our attention and implement it as soon as possible!

Thanks a lot for bringing this up!

[shub14]

Thanks a lot, really appreciate your efforts on this. I wanted to confirm will the Uint8Array changes help in improving the tool's performance, especially while loading large .mcap files?

[aneuwald-ctw]

Can you try out this branch? #586

I opened already a draft PR. I guess we should work on it a little bit, but if you can checkout to this branch and test, would be a nice feedback!

[shub14]

I tested on web as well as desktop applications. It has significantly improved on both but the web version is much faster than the desktop. Is there any difference between the configurations or optimization between the electron's engine and chrome?

[aneuwald-ctw]

Well, it should not have any difference in terms of data processing. Electron is a native app and has access to your OS, and maybe it gets a little slower for some reason, but it shouldn't be that different.