FilterHN

I still think we should find a way to integrate the two somehow - i.e. analyze locally and then send the pre-analyzed data for visualization purposes to the remote firefox profiler. Does anyone know a good format we could use for that purpose? It needs to be non-verbose to not hit the infamous 2GB/4GB JSON heap size limitation on import e.g. Similarly, we also need something that can deal with the various cost types we support in hotspot, most notably off-CPU time.

▲

pityJuke

10 months ago

[-]

samply [0] does something to send data to Firefox Profiler that I've had better results with compared to the regular import.

[0]: https://github.com/mstange/samply

▲

ognarb

10 months ago

[-]

From the same author, Heaptrack to investigate memory usage is also very helpful: https://invent.kde.org/sdk/heaptrack/

▲

throwaway0665

10 months ago

[-]

I just assumed they used the same gui toolkit. Awesome to hear they're from the same author.

▲

mananaysiempre

10 months ago

[-]

(The toolkit is Qt.)

▲

fransje26

10 months ago

[-]

Great tool that has been really helpful in helping me find unexpected bottlenecks in the codebases I've been working on.

It's easy to use, and pairs beautifully with the unintrusive perf tool, which makes the combination a joy to use.

And, if combined with a codebase opened in QtCreator, you can click on a hotspot in the flamegraph, and it will bring you automagically to the correct file and line in QtCreator, without any explicit linking required between the two programs. I discovered that feature accidentally, and the fact that it just worked seamlessly really impressed me. (Tested on a Debian-based Linux).

A big thanks to KDAB for making this tool available to us!

▲

milianw

10 months ago

[-]

You are welcome :)

And to people using other IDE/editors - you can configure which one gets opened when you click on a source line from the hotspot settings. QtCreator is just the default (when that is installed).

▲

kvemkon

10 months ago

[-]

It's a pity, Hotspot is abandoned in Debian. Just because of policykit [1]. The package seems to be rather easy to maintain (it has a tiny rule file, small control and only 2 patches).

[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1025573

▲

fransje26

10 months ago

[-]

And, of course, a big thanks to you for writing this tool! :-)

▲

directmusic

10 months ago

[-]

I primarily do macOS development work and use Instruments for profiling, but have been working on Linux a lot lately and Hotspot has been pulled out a lot over the last few weeks. Highly recommended.

▲

wizzledonker

10 months ago

[-]

This is a fantastic tool and I recommend it! I use it every day to recursively solve bottlenecks in our code base.

▲

pjmlp

10 months ago

[-]

Nice tool, thanks for the heads up.

▲

drewg123

10 months ago

[-]

I'm on FreeBSD, so its a lot of overhead to check myself: But, does this have a source view like Vtune, that can attribute cycles / cache misses / etc, to a line of code? I see it has an assembly view..

▲

dustbunny

10 months ago

[-]

I couldn't figure out how to use hotspot. I used speedscope instead: https://www.speedscope.app/

Maybe I should check out hotspot again.

▲

grandinj

10 months ago

[-]

Wonderful tool. Has been incredibly useful in optimising LibreOffice.