After effects quits, crashes or makes a blue screen error during rendering on one specific machine

[Tamas_Koszegi]

There is a problem which is driving me crazy for months now.

We have multiple computers running After Effects, and there is a brand new PC with no expense spared components in it which just crashes really often (Intel Core i9-13900K with 128 GB DDR5 RAM, RTX 3060 Ti and a MSI MAG Z790 TOMAHAWK WIFI) and a All other, much lower quality PC-s manage to continue working on the same comp, same AE version, same Windows, without problems. Its not a comp specific or effect specific thin, it can happen anywhere, even in  safe mode (both Windows and After Effects)

I have tried updating drivers, stress tested each hardware compoment separately including RAM, CPU and GPU, installed it clean, run in safe mode, deleted 3rd party plugins, uninstalled RAM modules one by one, ran without GPU, updated BIOS, checked Event Logs, analized dump files, used driver verifier, tried basically all solutions found online without a result. I am working on this with the support of ChatGPT for weeks without any solution.

Usually the blue screen messages are:

1. SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (7e)
2. DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
3. SYSTEM_SERVICE_EXCEPTION (3b)
4. IRQL_NOT_LESS_OR_EQUAL (a)
5. APC_INDEX_MISMATCH
6. ATTEMPTED_EXECUTE_OF_NOEXECUTE_MEMORY
7. UNEXPECTED_STORE_EXCEPTION
8. SYSTEM_THREAD_EXCEPTION_NOT_HANDLED
9. IRQL_NOT_LESS_OR_EQUAL

This shows into the direction of RAM / Driver problems, including memory access violations, potential driver issues, and system thread exceptions. I ran all RAM tests on earth and they all said everything is fine. Driver-wise I updated everything manually and with software help too.

Dump file analysis shows that the  faulting module is still within the kernel (ntkrnlmp.exe), pointing to a system-level issue. 

I just ran out of options and I have no idea what to do besides selling the computer piece by piece. Any ideas?

Dump files:
https://drive.google.com/drive/folders/164krg2Fd0UEiJrE3puBuMZuET3Gcuc1F?usp=sharing

[Tamas_Koszegi]

Oh you actually also have an Aorus Master + Vengeance combo, interesting!
Yes, my problem is exactly the same.

To me what helped a BIT was putting the Memory Enhancement Settings to RdEQ Mode 2, and then in Channels Standard Timing Control and Channels Advanced Timing Control I put everything from auto to its JEDEC value (this is just relevant to my Aorus comrade).

 It definitely has some revelance what kind of board do you have because my MSI Tomahawk crashes far less.
The e-core disabling provided me with conflicting results so far, it seems to me that I do have the same crash message than before, but when I say "ok" it just continues rendering 😄 But this is not backed up by enough testing, so take it with a grain of salt. Unfortunately hardware support points only to Adobe, and Adobe so far is unfortunately silent on the matter.

[jonathans94913674]

Did you try turning off the e-cores? Not ideal but it works for me as a workaround for now.

[Tamas_Koszegi]

Oh totally, when I read what you wrote I tried it out immediately, but I have a deadline now so I could not really experiment with any of the "crashing for sure" projects I have saved.  So far what I see that I still have the error message popping up sometime when I render in Media Encoder, but after I close it it just continues rendering. Strange. I will test it properly next week. I also have one last shot to buy an other CPU / mobo combo before it makes no more economic sense to buy more PCs (I had to buy an extra anyway), so I am thinking about completely switching to AMD, but I also heard horror stories about how that handles memory... I just dont know.

[AlexBalcombe]

In Bios I turned off e-cores and tweaked the Memory Enhancement Settings to RdEQ Mode 2, changing entries from Auto to the JEDEC values.

Still experiencing crashes and Blue Screens unfortunately.

Balint - you mentioned you have other workstations that run fine, mind sharing the highest performing one that is stable?

[Tamas_Koszegi]

I wish I would, but no. Both of my computers are crashing and it drives me crazy. I am actually thinking of moving away from AE just because of this bug, while I am an AE animation director for more than 10 years...  Anyway the computer which runs a bit more stable but still crashes is an intel i9 14900K with an MSI z790 Tomahawk Wifi (newest bios) and 192gb of G-SKILL Trident 6400 Mhz CL32 Trident Z5. It crashes less than the Aorus Master X with the Corsair Vengeance which crashes quite a lot, but its the same issue here as well. I would literally do anything to arrive to a solution.

[AlexBalcombe]

Thanks for sharing Balint.

Sorry your other PC's are buggy. For reference I still have my old computer; an i7-7700k, 1080Ti and MSI Z270 Motherboard. Very stable on that, just not as fast as current tech. Appreciate it wouldn't handle massive AE comps well either.

Never had an Apple workstation but seriously considering trading my current build for a Mac M2 studio as recommended by other editors I have spoken to. Would rid me of this curse for sure. Also due a laptop upgrade, and really hesitant to get anything with an i9 in it now.

I wouldn't let this be the thing that drives you away from your career - a lot easier to get a new PC than a new job these days.

[AlexBalcombe]

Hey Balint,

I have tinkered with my BIOS settings and my problem is now solved. I have stability and zero crashes (so far) across multiple After Effects and Premiere projects that were 100% guaranteed to crash previously.

There is clearly an issue when using i9 13900k or 14900k CPU's with Z790 gaming motherboards, despite all blue screens and error logs pointing to memory. The 'auto' BIOS settings are not suitable for a monster turbo powered CPU.

The instability from my CPU was the Turbo boost. Enabling E-cores is fine for me, but try keeping them off first. I've also made sure to keep the RAM within reasonable limits, although so far I'm not sure if any of the RAM settings are actually contributing to the success here . Try this...

Go into BIOS, and:

- Set Gigabyte Perfdrive to Spec Enhance (or E-core disable to be extra safe)

- Set your performance CPU clock ratio to 57 or lower (mine is at 54) and Efficiency CPU clock ratio to 44 or lower.

- CPU/PCH Voltage control: CPU Vcore as 'normal'

- In Advanced CPU options, set Loadline Calibration and Vcore/Overcurrent protection to Medium

- Disable CPU EIST, Race To Halt, Energy Efficient Turbo, Turbo Boost, Turbo Boost Max 3.0, Flex Ratio Override, and C-states

- Make RAM XMP 5600MHZ or lower (mine is at 4800) and disable any RAM 'booster' settings

- Go into your Advanced RAM/ SPD settings and use all the JEDEC values. Same for Channels Standard Timing Control and Advanced Timing control.

Other things that were recommended regarding this problem, but I didn't need to do myself:

- Disable multi-core enhancement in CPU options

- Set turbo limit to "Intel POR"

- Set SVID behaviour to "Intel Fail Safe"

- Reducing power settings for Long and short duration power limits

- Setting power limits in Intel XTU to 250W

Really hope this helps.

[Tamas_Koszegi]

Okay so first of all, this is a problem which has been bugging me for MONTHS, and I have to offer my most sincere gratitude to you sir, who managed to finally solve it after so many sleepless nights. Thank you!!!!! Yes, it solved it all. I am so happy. Thank you!!!

Now with that said this is a huge problem, because disabling turbo boost does massively take back from performance. Do you think this is a CPU issue which can be solved with RMA, or just all Intel CPU-s are like that? Because I did RMA my 13th gen i9 and Intel paid it back fully (so they do think they have a part to play in this), but now this is my SECOND i9, seriously would this be faulty AGAIN? I just have difficulty believeing this. So while finally finally my computer does not crash (thanks again) but I dont think this issue is over, I need to investigate this further to arrive to a better conclusion. Do you think changing to Ryzen would help?

[AlexBalcombe]

Hey Balint,

So glad we've both fixed the issue.

I believe Intel are aware of this issue between CPU and Motherboard - whether it's something that can be fixed by getting a new one I do not know. Considering you already did an RMA I think maybe not.

Performance wise, as long as you set the 'Performance CPU Clock Ratio' to 58 (5.8Ghz) and 'Efficiency Clock Ratio' to 44 (4.4ghz), you theoretically should not see a change in performance. The removal of Turbo probably just affects energy efficiency in some way.

[jonathans94913674]

I'll have to give this a try as well!

[Tamas_Koszegi]

I am not that convinced about the performance drop, but we will see, I will run some tests when I can. One thing for sure is that I never had a crash since. Also there is a huge upside knowing that the turbo boost is the problem because there is a huge amount of knowledge online about the turbo boost issues, I just never ever connected the 2 things together. So thanks again, I will test it with both of my computers and I will come back with the results!

[Tamas_Koszegi]

So when I thought this thread is coming to a close a new twist in the events came up 🙂

Fortified with the knowledge from AlexBalcombe I bought an OTHER i9 14900K to put into my assistant's computer. However, just out of curiosity, I installed it into my computer first. I left everything as it was, all hardware, software, I just swapped out the CPU, and I cleared CMOS to see if the computer renders with the original, crashing settings. And to my surprise it NEVER crashed even once. I was torturing it all night with all the guaranteed for crashing comps, and NOTHING. 

This sheds a new light to this issue… it seems like it has a lot to do with the dreaded “silicon lottery”, where there are subtle differences between chips. (in this case, the chips IMC)  An other proof to this was that my first crashing CPU was RMAd by Intel without any issues whatsoever. And yes, it seems like I was unlucky enough to buy 2 faulty CPUs after one an other, but finally this one seems to work without any issues whatsoever. This issue does not seem to be Adobe's fault, it looks more and more like Intel's fault. I strongly recommend anyone who faces this issue to RMA the CPU, because there is a good chance that the second one will be better! And if you can not do that, Alex Balcombe’s solution seems to work like a charm. 

[jdmcclintic]

Take my upvote!!! Tried the equavalent (Turbo Off, Clock Ratio of 54 and 44) on my AS-Rock 790 MoBo 14900kf combo, and so far a Blender render that was crashing is finally rendering. Now to bring that into AE and see if it crashed it or you have performed a miracle that has already cost me 7 days or work. I can't complain if it is a little slower, as I was coming from an ancient 5620k i7 (6 Core). 

[Zombie_z1742]

hey thanks for sharing I was having the same problem and didn't know why but damn it was becoming so hard to work for me and I was in the middle of a project. I've just applied the settings you recommend and now I'm testing it. so far it hasn't crash but let's see if it stays like this.|
I've 14900k

aorus pro x rev 0.1  and bios f3
32gb 6400mts corsaire vengence
in the longer run shoud I sell this pc for something more stable like I'm thinking about selling this machine and buying a mac m2 max cause I was using that for my office work and that laptop is exceptionally good.

[jdmcclintic]

Step one for the i9 crashes is to update your BIOS firmware. At this point, most MoBo manufacturers have gotten the guidelines from Intel on several fixes, the latest of which started rolling out this month. You may still have to dial in the clock multipliers depending on how lucky you were with your chip, and whether or not it sustained damage from extreme voltage from the MB Bios that came with many systems. 

[onnpro87]

I have a brand new HP Z2G9.. very expensive. total garbage. 
13900k
128GB DDR5
RTX A5000
exactly same issues you described here..
HP replaced the MB (2 times!) RAM, GPU, SSD and CPU.. same problem.. I really can't work anymore on this machine..
i'm working for the last few months on a older laptop with no crashes at all..
I really don't know what to do.. I thinking about buying a new custom PC but afraid the issues will be the same

[Tamas_Koszegi]

I am very sorry to hear this, welcome to the club! Since we arrived to a (kinda) solution in this thread I read multiple articles saying that the number of faulty i9-s is bigger than any returns before with Intel. This affects 13th and 14th gen (as they are basically the same) I would still press for CPU replacement is you can, or if you can not, try out Alex Balcomes solution because that is still the best fix we found. If not... I would sell it. I am waiting for my second RMA now, if they wont exchange it I will sell it, and then buy from a place where they accept returns, just buy 3, try them all and return the rest.

[onnpro87]

And of course HP Z workstations BIOS is very limited, so I can't dial all the configurations people suggested here.. I'm stuck

[onnpro87]

HP replaced me 2 motherborads and 2 CPUs.. it is not a faulty unit. it is a very poor designing and manufator of CPU which you have to cut off and limit on your BIOS (which I can't do on my machine)

[Tamas_Koszegi]

Ahhhh okay that is a problem. Have you tried disabling Turbo Boost in Windows? There are tutorials online saying it can (kinda) be done.

https://notebooktalk.net/topic/464-quickly-turn-turbo-boost-on-or-off-in-windows/

[onnpro87]

Do you think it will fix it? I'm so tired.. I got 5 HP technician visits here on the last 2 months.. every time I need to reinstall the system just to figure that I got the same issue... My PC in unplugged for 2 weeks and I don't have time / want to connect it once again until I found a soultuion.. I just wrote HP and email describing what I should change on the BIOS.. maybe there is an expert mode that can open for me so I could check..
Thanks to your thread I just cancelled custom $6000 new PC I ordered with same components.. I'm so glad I saw this post, otherwise I would have here 2 useless computers..

For me is waiting for a solution from HP or switching to MAC (which is also not an easy solution, because you can't use the proImport AAF on MAC anymore..and I'm using Avid AAF daily..) Ahhhh! before this 13th Gen everything was just great

[Tamas_Koszegi]

I am extremely sorry 😞 If that is any consolation I know how you feel, I was plagued with this problem for more than a year, one night before a big deadline I was actually crying in front of the computer, begging it to finish rendering. Thats why I made this thread because I thought I can not be the only one out there.
Unfortunately nothing I wrote is official help, I am not a PC expert by any means, I am an animation director who has a bit bigger test pool than others because we have multiple PCs in our studio and I could try and switch out components. The only thing I know is that the turbo boost switchoff was the first time when I actually saw replicable and testable results. Switching RAM units and giving them all kinds of timings made very minimal progress, switching off the e-cores gave it some progress but it still crashed, the turbo was what could be repeated with the same result over and over. Yes, the BIOS settings here are quite extensive but most of them is mostly for safety and to disable some of the automatic boosting features of the Aorus board, I am pretty sure you dont need them all.  Anyway, tell us what happened and good luck!

[onnpro87 · Accepted solution]

You will not beilieve! IT WORKS! I only changed the CPU maximum power to 99% and edit the regedit to open the performance boost mode, changed there to Efficient Aggerssive.. I guess it is doing a similar job to the limitation you did on BIOS.

Since yesterday the computer works just fine. unbelievable.

Intel have a serious fault on their 13th and 14th cpus.. 

[Tamas_Koszegi]

I am so happy it works! See, power of online communities 🙂 I hope this thread can help other people as well!

Yes, Intel has some serious problems going on with Raptor Lake / Refresh.  They are aware too as they do exchange them. Still, it is possible to find "golden samples" that do work, as I managed to grab 1. Now I am on the hunt for an other. Still, its just ridiculous that we industry professionals are just tortured by all these bugs. In the meantime, can you give an estimate about how much performance drop do you see? Because that is still debated. Happy animating in the meantime!

[onnpro87]

This is so frustrating that we have to do our daily job and to be experts
ITs..
But yes, the power of the online communities solved an issue that 5 HP
technicians couldn't solve.
I can't tell exactly how much it changed, but I always work with the task
manager and I see it doesnt go above 5ghz (instead of 5.8ghz it should..).
Anyhow, it is a big upgrade from my laptop.. At least I can work with my
128GB and the RTX A5000..
Thanks again for all your help!
--
*Onn Henner*
*www.ohpro.net <> *
* onn@ohpro.net | +972 (52) 8478041*
Video Editor* |* Motion Graphics Designer* | *Photographer