13:15 <yyp> Is there a way to profile river? Today it hit 50% CPU load and almost froze the computer, now I wonder why did that even happen

13:19 <ifreund> running with the DRM backend I'm not sure, tracy might work

13:20 <ifreund> running nested you can use pretty much any normal profiling tool without issue

13:20 <ifreund> I guess perf record would work just fine on the drm backend as well

13:21 <ifreund> yyp: were you doing anything unusal that might have caused that to happen?

13:22 <yyp> No, just doing casual work

13:23 <yyp> Anyway, I'll try tracy

13:26 <ifreund> yyp: see https://github.com/ziglang/zig/blob/master/src/tracy.zig for basic integration

14:03 <travankor> yyp: could it also be lack of damage tracking?

14:04 <travankor> if it only happens on DRM backend, then that is probably it iiuc

22:29 <leon-p> late to the party,but does valgrind work with zig code? If yes, I find callgrind is a nice profiler, thanks to the call tree you can display with kcachegrind. river will run pretty slow like that, but if you can find out the problematic task before hand and do a "lab run" it is pretty handy.

22:30 <ifreund> valgrind works great with zig code :)

22:30 <leon-p> nice!

22:31 <leon-p> probably my favourite debugging tool

22:31 <ifreund> zig's `undefined` also has valgrind integration

22:31 <ifreund> so you get really good diagnostics

22:31 <leon-p> I'll have to try that at some point