Unreal Editor 4.9.1 hangs with 101% CPU time in activity monitor. This occurs quite often though. I couldn’t find a specific reproduce steps.
Environments:
Took some screenshots for additional info.
http://d.pr/i/1las0
http://d.pr/i/JgpV
Thanks!
Hi ,
Please sample the editor in Activity Monitor by opening the Activity Monitory, selecting the editor in the list, using View->Sample Process to generate a report on the editors current activity, and saving this as a text file. Then post a copy of the text file here.
Thanks
I cannot encounter the issue anymore. I will report back with process sampling when this occurs again. Thanks.
Hi ,
I’m marking this resolved for now. If this issue returns, simply respond to this post with the process sampling. -Thanks for the update.
The Activity Monitor samples did not turn up anything unusual so we will need some more information. Do you have other versions of the Engine installed? If so, do they also exhibit this lag? (If not, can you download a previous version and test it?) Also, does this happen if you open an empty project template, or only with projects containing your own assets? If only your own project or projects, would you consider these large with lots of assets or smaller with fewer assets?
Also, are you running the latest iteration of the Engine, 4.9.2? Any and all information you can provide will be helpful.
Now, I switched to the default driver to see if I can reproduce this. I will report back if I encounter once more. The issue seems to occur very randomly.
Thanks!
Hi ,
Haven’t heard back from you in a while so I am marking this thread as “resolved” for now. If you continue to experience this lagging, simply respond to this post and we will continue to investigate. -Thanks.
,
I encountered this issue again when importing an “ao” texture in png. I sampled the process again, just in case. I also observed “import” button is still highlighted yellow even after file open dialog disappears.
Let me also try 4.10 preview 1 if I can encounter the issue.
Thank you.
I’m trying 4.10 preview 1. I don’t seem to have this issue in the preview version. Just making a note here. Thanks.
Hi ,
I discovered the same freeze in 4.10 preview. It happened about five times over 5 hour work time. I spent some time to compare process samples I gathered so far.
I made a note in this link:
Unreal smart pointer impl might synchronously releases ref count. Something else is waiting for that thread to finish as well.
Hi ,
You mentioned your project is fairly small, but have you tried this with blank templates like the Third Person or First Person Template, migrating in a few of your custom assets in at a time? It may be a custom asset that is causing this conflict.
I can try that.
However, I also encountered this without importing as well. I used Perforce control and, when I checking files into the depot, the editor asked “This takes too long, do you want to continue?” with “Continue” and “Stop” buttons. At that time, the editor entered into the same state with the same “leaf” of call stack when process is sampled.
Anyway, I will try to do incremental imports when I get a chance most definitely.
Also, since I’m just working on a personal project, I’m totally fine sharing the entire project if it helps.
I get a “Page no found” for the droplr link you’ve provided, but a copy of your project would be helpful. You can try that again and/or send a dropbox link via private message to me on our : https://.unrealengine.com/index.php
Hi, I set link expiration to one hour. Sorry about the confusion.
After further investigation into the issue, and inspecting source code, and googling about _AMDDeviceDetached event generated from Mac OS X,
I discovered a reliable way to reproduce this issue.
After source code inspection, I also found that my earlier guess was close. 
At line #756 in MessageEndpoint.h, SafeRelease enters infinite while loop. It indicates something else is not releasing FMessageEndPoint object.
For full story, my iOS devices were connected to a USB hub; then, the USB hub was connected to my iMac. The hub seems to be unreliable as it reconnects iOS devices every so often. In my case, once an hour, that’s pretty much why it froze my work once every hour.
I’m working again after connecting the device cable directly into iMac now to see if the problem disappears.
EDIT: I just encountered the same freeze when the device cable is directly connected to my iMac. It seems the sporadic disconnect events are usual for Mac. I will try to disconnect the device completely and continue the work to reproduce the issue again.
Another update.
I encountered the freeze without even attaching the device at all. http://d.pr/i/1fehJ It happened as soon as source control related dialog appears.
As shown in the process sample at http://d.pr/f/FTBJ , _AMDDeviceDetached event occurs without having a mobile device attached at all. It ends up with the same call chain and enters the infinite loop.
The replication step mentioned before was just one case which can cause _AMDDeviceDetached event, not all. Whenever _AMDDeviceDetached event occurs, the editor hangs.
Hi ,
Thank you for providing all the pertinent information to reproduce the issue on our end. This is currently being investigated as a bug reported as JIRA [UE-22976]. Should this issue be resoled in a later release of the engine, we will notify you with an update to this post.
Again, thank you bringing this issue to our attention,
.
This issue is being looked at under a new JIRA [UE-25123] that includes additional data from other users. The original JIRA, [UE-22976] has been closed. When this new bug report is resolved, I have added a note to update this post to to let you know.
Update: JIRA [UE-25123] has been marked resolved for a future release of the Engine.