Random stutter (high cpu usage), returns to normal after exit and re-enter map

reference number 3280045984

The game will drop to 10 fps from 100fps during random mapping. This started this league. It will trigger the mangled text bug in early league. The stutter will stop after clearing out some off screen mobs, or waiting for long period of time. It also stops if you exit and re-enter the map

However, this bug still exists, but it no longer gives the mangled text after the stuttering.

Here's my recording on youtube.
https://youtu.be/vxOp4E7udKk

You can clearly see the cpu time is very high and low fps. The cpu time and fps returns to normal upon exit and re-entering the map.

It also causes high current draw from the power supply since the lights in the room starts flickering as well.

This is still happening. Multiple times a day. Sometimes the lag disappears after few seconds, sometimes I have to leave the instance to fix it.
I would assume that there is bug in client code that consumes cpu or gpu resources. It started about a 3/4 year ago.
Even my friend with 3080 GPU has this issue.
I have 1660S, i5-9600KF, using Vulkan.
My FPS goes from 60-120 to 10, even when nothing is on screen.
Leaving the instance always resolves the issue. It is not game breaking, but annoying.
It happens in hideouts, cities, everywhere.

Got the same problem, almost indentical what you show on recording. FPS randomly drops from ~120 to ~60 with extreme stuttering.

RX 6700 XT
Ryzen 9 3900X

I gues we can only hope GGG will fix it.

Same problem here. If I remember correctly, for me this is started to occur after the 3.20.1b Patch. I have tried several common fixes and this problem has not been corrected. Weirdly Alt-tabbing to another window immediately fixes the problem, so while this is not critical but it is annoying.

OS: Windows 10
CPU: Ryzen 7 5800X
GPU: RTX3070

No problems on other games, hope GGG will fix it.

I have discovered that turning v-sync off and on also fixes the stutter. So we do not need to leave the instance.