Processors
Intel® Processors, Tools, and Utilities
15678 Discussions

I9-13900K and I9-14900K Memory Access Violation with turbo

Adrien73
Beginner
8,124 Views

Hi,

 

We found out on github our issue related to playing a specific minecraft content was our CPU turbo speed causing Memory Access Violation.

Here is the github link : https://github.com/FTBTeam/FTB-Modpack-Issues/issues/4706

 

Does this kind of issue have any chance in the future to be solved by an intel driver update or a bios update ?

 

Docker command to test the issue : 

docker run -it eclipse-temurin:17.0.10_7-jre-jammy /bin/sh -c "cd /home && curl -JLO "https://api.modpacks.ch/public/modpack/117/11701/server/linux" && chmod +x serverinstall_117_11701 && ./serverinstall_117_11701 --latest --auto && echo "eula=true" > eula.txt&& chmod +x start.sh && ./start.sh"

Append on both linux and windows.

On linux it does a SIGSEGV (segmentation fault) and on windows a memory access violation.

 

Limiting the CPU speed to 5Ghz fix the issue, but I don't see the point to buy those CPU if you have to use them below default specs.

0 Kudos
7 Replies
RamyerM_Intel
Moderator
8,049 Views

Hello Adrien73


I need some more information in order to help you. Would you be able to share the SSU logs of both of your system so we can further check this for you? Also, what is the Minecraft version that you are referring to? Does this issue also happens with other games? Additionally, upon checking, both the error you received on Linux and Windows is related to illegal memory allocation. May I please know the RAM that you used as well on both your systems?


Ramyer M. 

Intel Customer Support Technician 


0 Kudos
Adrien73
Beginner
8,004 Views

Hi,

 

Thank you for your interest.
Here are the informations you asked for :

I only have one system, I just tried on both native windows and through a WSL2 docker container to exclude some windows config related issues (even if this solution isn't as good as a second system).

 

The minecraft version is 1.19.2, and the versions of the modpack tried were between1.0 and 1.1.
The issue only happend with this specific modpack as with the same docker command I managed to host other packs with no issues.

 

The exact ram model is 2x16GB DDR5(CMK32GX5M2B5200C40) 4800Mhz

 

To be clear, the only reason why I'm posting here and not on a more minecraft related forum, is because the only people (4) who claimed having this issue were running a i9-13900 or i9-14900 processor, and the solution to this was to limit the CPU frequency to 5Ghz in windows power management.

0 Kudos
lokran
Beginner
7,989 Views

Sorry to hijack this, but interesting, as I just RMA'd my 14900K, which was suddenly giving me status access violations in Chrome and other light apps as Discord and keeps on crashing. 

 

Cooling is totally fine, everything BIOS wise stock, of course no XMP, as this is considered overclocking which is forbidden.

 

RAM seems fine, also thinking it is CPU related. But not willing to clock down, as this is supposed to work with the clocks advertised and paid for. 

 

 

0 Kudos
RamyerM_Intel
Moderator
7,790 Views

Hello Adrien73, 


We sincerely apologize if it took us some time to respond. Upon coordinating with my team, we kindly ask for you try the following troubleshooting steps.  


On the previous setting you mentioned, kindly please run XTU and collect the following logs. (package temp., CPU utilization, Package TDP, Core TDP). You may download Intel XTU at this link: https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-xtu.html. You may also send the HWInfo log with 500ms step. 


After doing so, we want to check the performance of your default BIOS. Kindly please try the following steps: 

  1. Set the BIOS to default
  2. set 253W in turbo boost power max and turbo boost short power max 
  3. set IccMax to 307A
  4. run XTU and log to file (package temp., CPU utilization, Package TDP, Core TDP),
  5. gather HWInfo log with 500ms step
  6. After doing so, test performance with AVX1 or AVX2 test. You may also conduct CPU stress test from XTU or any default apps intended for your system. 

If these parameters shows no issues in your system, you can set these parameters in the BIOS. 


Please feel free to forward us the results. Furthermore, we would like to manage expectations regarding third-party applications. If you encounter any issues that are not yet compatible with Windows 11 or our CPU architecture, we encourage you to reach out to the app developers directly to address the required updates.


As for lokran, we want to give you the full focus of our support, kindly please create a new thread so we can try to resolve the issue with your system as well.


Ramyer M. 

Intel Customer Support Technician 



0 Kudos
Adrien73
Beginner
7,763 Views

Hi,

 

Thank's for your answer, I have followed the steps you asked.

 

First of all, I don't ask you to find me a solution, I found 2 solution (One today, more at the end), and the solution I used up to now was to limit in windows power usage setting the CPU speed to 5Ghz like I mentioned earlier.
I post here mostly to try to help other people that could encounter an issue like that, or even you and because the reports we gathered were all coming from the same CPU models as the title stated.

If you firmly believe the issue is more related to the process running I think we can close this post without further research.

 

First of all, all the time my BIOS was set to default, as before "identifying" the issue was between the process and the CPU, I tried to tweak my ram, but my last try forced me to reset the bios by removing the CMOS. It was confirmed when I got in the bios today, and it said "Bios set to default after CMOS removed" or a message like that.

 

I did all your step, but to be able to run the XTU software I had to disable "Virtualization Based Security", I don't know what it means, and I indeed in the past used Hyper-v (or close to this) for software like docker.

After disabling it, I tried to run the process I came here for, and it crashed exactly as before.

 

I did the tweak related to the Turbo Boost Power Max, Turbo Boost Short Power Max and Processor Core IccMax in the intel XTU tool.

 

Before running the test, I disabled the power management clock limit of 5 Ghz that I was using since I posted here, to be able to run my game without crash.

Then I ran the test with both software logs enabled as asked, the logs are around 12 minute long, and I ran, again from the XTU Tool, AVX 5 minute test, few seconds after it ended AVX2 for 5 minute, and then I ran the java process that crashed before and was the reason I posted here.

 

When I ran the java process it didn't crashed.
It was unexpected for me, so I tried changing the option you asked me to set one by one to identify which was fixing the issue.
It was the Processor Core IccMax that was also fixing the issue, I don't really know what it means, and if it's worse than limiting the clock speed in windows, but I checked and the process was now not crashing, and running over 5Ghz (Around 5.5Ghz almost constantly according to XTU).

 

Again, thank you for your interest in this issue,

Adrien.

 

Edit : Did some test to find up to what value the Processor Core IccMax was working, V mean it worked, X mean it crashed, ordered by my test order.

307A V
327A V
347A V
367A X
360A X
357A X
352A V
355A X
354A X
353A V

May this information help in any way.
Edit 2 : This test was fine for the generation of a new world, but I had to lower it even lower to load an existing world, I'll now use 335A that seemed stable even for an existing world.

0 Kudos
lokran
Beginner
7,743 Views

Thank you for offering help!

 

Though I received a new 14900K after just one day and now system is totally stable and no more "status access violation". Nothing else changed, just switched the CPU. So I must have had a defective core or IMC on my first CPU unfortunately. Never happened to me before, as Intel is very reliable usually!

0 Kudos
RamyerM_Intel
Moderator
7,621 Views

Hello Adrien73, 


Thank you for clarifying. It is wonderful to see that you aim to help our community. Thank you for trying our recommendation and we will be taking note of this feedback for now.


As for lokran, 

I am glad you are no longer experiencing the issue. In case you need further assistance in the future, feel free to create a new post as this thread will no longer be monitored. 


Thank you both for engaging in our community forum. We wish you the best on your future endeavors. 


Ramyer M. 

Intel Customer Support Technician 


0 Kudos
Reply