I noticed that ConanSandbox.exe isn’t always responding using Task Manager. I disabled Isle of Siptah a while back which means I have no mods and no active DLCs, just the base game. I verified the Conan files and confirmed that Conan, Steam, Nvidia, and Windows are all up to date. I saw a suggestion online to right click on the Conan application file, run as administrator, and see if I get an error message.
I got the following message:
Couldn’t start:
“C:\Program Files (x86)\steamapps\common\Conan Exiles\ConanSandbox\Binaries\Win64\ConanSandbox-Win64-Shipping.exe”
CreateProcess() returned 2.
I don’t see a ConanSandbox-Win64-Shipping.exe file in the suggested location. Am I missing a file? Or is this a file that is supposed to be created but isn’t created? Not sure what to do with this error message.
Ok I got told the wrong executable. The correct executable is C:\Program Files (x86)\steamapps\common\Conan Exiles\ConanSandbox\Binaries\Win64\ConanSandbox.exe
So the question is still how can I fix ConanSandbox.exe when it stops responding periodically?
I wouldn’t use steam libraries. The problem is usually the file name length. There is some (text) savings in resolving those issues. I wouldn’t use winehq at this point in time.
Fortunately, Microsoft did publish their own CRTL. It works, but that probably means nothing to non-devs. I like MinGW. It is a nice platform. It works. LL’s auto-build… not worried about it if it gets wrapped up in one of those platforms. I think dueling autobuild with .NET (core and/or framework) may provide quicker insight.
It’s been a while since I used auto-build. It has no business on my workstation. I’m just trying to adjust the matrix hardware to not be so virtualized. Those aspects are costly. It swaps the state of the matrix onto the stack, which may be suboptimal. I’ll have to RT*M on the matrix to make it more like a job queue with streaming support. I compared Xeon and i5 for it’s fast ram, and Xeon may be more parallel. There is no locution either way. A kid could write the code and show the speed of the i5.
From your post and example you are using windows, @Dzonatas is talking about Linux when he is referring to wine, which does not seem relevant in your case.
Since either the last update or previous update when I exit the game it does not always respond and I also have to use task manager to close it.
It seems as if you are getting this at runtime according to the section of log you have loaded, there can be a multitude of reasons as to why this is happening and more information would be required to try to assist you in troubleshooting this.
The “not responding” status in the Task Manager is determined via a heuristic that the Task Manager uses to measure the responsiveness of the UI*. So the question I would have for you is: regardless of what the Task Manager says, does the game respond to your input?
* For anyone interested about the source of this claim, it’s from a Stack Overflow comment by Raymond Chen, the author of The Old New Thing blog and an expert in the obscure internals of the Windows operating system.
When its acting up, I can hold down E to walk forward. I will step and stop repeatedly. A clanmate can walk beside me and see me stepping and stopping. That same clanmate will get to the destination about twice as fast.
I came across his blog the other day with regards to walking the stack, I had never thought to use tasklist via cmd let alone realise how many switches there are.
Thanks for the detailed description of the symptoms. From what you described, it’s clear that it’s not the Task Manager’s heuristic that failed, the game is actually being unresponsive in bursts.
The next step would be to try to figure out what’s going on. Is it being unresponsive because it’s starving the CPU and cannot schedule the input handling properly? Or is it something else, like some intermittent problem in receiving inputs?
For that, it would be helpful to observe the system resource usage – CPU, GPU, memory, I/O – both for the game process itself and the whole system. I would recommend switching from Task Manager to Process Explorer. To make sure you can see graphs for all system resources, after you open Process Explorer, you should go to “File” menu, then select “Show Details for All Processes”. To see the graphs, you can press Ctrl+I (or go to “View” menu and select “System Information”). To see the resource usage for each process, you can go to “View” menu, then click on “Select Columns” and then make sure that the following columns are visible:
“CPU Usage” from the “Process Performance” tab
“Private Bytes” and “Working Set” from the “Process Memory” tab
“GPU Usage” from the “Process GPU” tab
“Delta Reads”, “Delta Writes”, and “Delta Total Bytes” from the “Process I/O” tab
With these things, you can see if anything looks abnormal when the game starts to struggle.
Right when it seems to struggle I see a huge spike in outbound Internet traffic and the executable becomes unresponsive. When I step and stop, there are a bunch of spikes on Ethernet.
When the game isn’t struggling, my Ethernet usage stays pretty low and doesn’t have the spikes or periods of unresponsiveness.
Mods… need an indeterminate amount of time, but the mod coders could use the other 60% for responsiveness. Surely there is a good way to do that… I can help with j2k if needed.
Best I can tell by standing in an area without any buildings and few NPCs, the game seems to quit responding to my outbound traffic and then there is a spike of incoming traffic when it starts responding again. The incoming spikes seem to be going from 24ish Kbps up to about 120-500 Kbps.
I can confirm that the antivirus and firewall aren’t blocking anything. I tried disabling IPv6 and the amount of Internet traffic reduced all of the way around. Unfortunately that didn’t resolve the issue.
See, it would be better if all those IPv4 drones had a home in that region… sometimes you’ll have to contact the ISP to expire past licensing. It’s rare.
The traffic spikes are definitely a symptom, but I wonder if they’re the root cause or if something else is causing the problem and the traffic spikes are just another manifestation of that.
From what I know of how Conan Exiles works under the hood, traffic congestion is more likely to cause rubber-banding than intermittent lack of response. The game should still be able to receive your input and simulate the behavior on the client-side until/unless it receives a correction from the server. That’s often what causes rubber-banding: the game starts moving you in the direction you’re supposed to move and also sends that same input to the server, but due to the time it takes for that input to send to the server, the responses from the server are based on the simulation without that input and you get yanked back to where the server says you “should” be.
That said, the traffic spikes might not be due to congestion, but rather some other problem. This is just speculation on my part, but if something in the netcode managed to somehow block the execution of other logic in the game, that could cause the problems you’re seeing. I don’t think that’s likely, because Conan Exiles, like most games, uses UDP for communication and as far as I know, there shouldn’t be any blocking logic in the netcode.
Have you tried playing in single-player to see if the same glitchy behavior happens?
It has no business on my workstation. I’m just trying to adjust the matrix hardware to not be so virtualized. Those aspects are costly. It swaps the state of the matrix onto the stack, which may be suboptimal. I’ll have to RT*M on the matrix to make it more like a job queue with streaming support. I compared Xeon and i5 for it’s fast ram, and Xeon may be more parallel. There is no locution either way. A kid could write the code and show the speed of the i5.
