VEAI shuts down the computer

Topaz Video AI General
8

I think what may be happening is a spike in temperature on the CPU or GPU. I know this can cause the computer to shut down abruptly. The reason I think this could be it is because I kept Open Hardware Monitor on during an upscale. On average I was getting temperatures on my CPU around 83 and this went on for hours. When I came back hours later though, my temps where still averaging around 83 but I noticed that the Max temperature recorded was 106! Something made my PC temp spike to 106 even though I had left it alone with nothing but VEAI running. (I did have 3 instances running. Two on my 3090 and 1 on my 2080Ti. CPU is a Threadripper 3970x 32-core, 64 GB RAM.)

I may try to lower the use of my CPU and VRAM in the settings and try to see if I get this spike again.

1 Like
9

okay but this should not be an issue with video enhance itself. maybe check the cooling options in the bios. i also think the cpu should normally clock down or start to throttle before it shuts down for safety. also check the maximum power consumption needed under full load and if your power supply can deliver that.

10

I am experiencing the same.
Because of still “low” usage of the RTX3090 GPU Power, i try to run at least 2 instances, sometimes 3.
Right now i am running a single VEAI instance, 1080p to 1080p Artemis-MQ 12(cleaning/smoothing) with a GPU load of 7% … CPU load is 30-50% , spf is around 0.15
GPU and CPU temperatures stay below 70°C

That “funny” 7% GPU load is strange.
I see that since V2.1.0.1b and after installing the latest nvidia studio driver v461.92

So i can easily run 3 instances
But running more than 2 instances “sometimes” leads to a crash.
“Maybe” caused by temperature peaks above 100°C

As some devs recommended to run multiple instances, at least until the RTX30x0 GPUs are better supported = used to the max. i wonder why there is no progress. i always hope for the better when i install studio driver updates, but no advance , yet.

I often requested a “saving queue” option. Like every other video software has for saving projects.
A large queue file (a list of 10 videos to be rendered) is a PROJECT for me and sometimes i spent lots of time preparing these queues.
During a crash … its all lost. Overwritten by zeros… the auto saved save.json is empty and i need to start from the beginning. No chance to load it again.

So running multiple instances has another backdraw besides the possible “crashes”.
The “save.json” file from one instance is always overwritten by the “save.json” file of the next instance.

But no answer, yet.

1 Like
11

Look into overheating. Most CPUs have protective circuits in them that either reduce performance or crash when the temperature goes over the max for that device. My Z490 machine with a powerful i9 processor in it kept crashing, because of this. I have very good water-cooling but it appears that the core voltage was the problem. I lowered it to the default using the Intel XTU utility and the problem went away. This also resolved a similar issue in another high-end rendering application I use.

12

Overheating / too high of power draw is usually the culprit. A brief spurt in power consumption can make a lot of lesser quality PSUs and Motherboards give out. There’s a reason why when you’re buying motherboards, you can pick ones with anywhere from 4 to 16 VRMs. If you’re going to be doing intense work, get equipment that can take that intensity.

Example:
[Removed]

For my Desktop, I got a PSU that had no measurable ripple under heavy loads - which means that the testing equipment of the reviewer needs to be more sensitive, and the PSU is ridiculously good at dealing with spikes in power consumption. (A Seasonic Prime Ultra Titanium PSU, for anyone wondering.)

I use ThrottleStop on my laptop to fix a similar overheating/overvoltage issue.
https://www.techpowerup.com/download/techpowerup-throttlestop/

A good old -60mv undervolt on the CPU and IGP did the trick.

1 Like
13

I have a similar story.
I believe it is cayse of one of the Windows 10 updates and Video Enhance Aı cannot use system resourses after the update. I wrote to support and send a couple of diagnostics. Now they are investigating. Here is the details:
I am pretty sure that version 1.5.3 to 1.7.1 quite good and was working on my PC (CPU: I7-8700, GPU: MSI RX 5700 XT Gaming, RAM: 32GB-4000mhz, Samsung 980 Pro nvme). Even increase the GPU slider up to 80 was fine for the software.

When I want to try 2.1.1 it starts crashing my PC in 3-5mins no matter what settings/video file type I use. After my 2 day marathon to figured it out what is wrong, I reinstall fresh Windows to both of my machines.Then tried all af the previous versions (v1.5.3 / v1.6.0 / v1.6.1 / v1.8.1 / v1.8.2 / v1.9.0 / v2.0.0 / v2.1.1). All of them crahsing.

I changed GPU to another slot and I tried another video file. I cleaned the video file with Premiere Pro and export as xmf. As I wanted the color correction after VEAI enhance.
The xmf file is 64GB and 1280x720. I loaded to VEAI and choose to HD 13920x1080 mov file. My PC crashes after 1/4 of the file. BTW: When I was looking through web found so many users have problem with crashes. I came accross an review about VEAI which saying is the software file size limitation is 10GB. But review was quite some time ago. I don’t know it still has that limitation or not.

I watched of all process, tried with smaller file size and figured out VEAI not write data to disk during process. I think All data kept in either memory or temp file. Don’t know it is related to it.
I am an an IT expert but I am out of options. Same PC, same specs.GPU checked, PSU checked (voltages seems ok), CPU checked, slots ok. Other hungry resource apps (4 games working well with 1920x1080 resolution). If the software was working with the version 1.7.1 previously why doesn’t work now. Only two things changed/added to PC: Windows 10 Updates on March and April. Seems that the software not compatible for the latest updates of Windows 10 (I don’t know). Seems like VEAI cannot handle for using the system resources well. Because all of my Adobe programs, my resource hungry games working like a charm.

Here is my another input to the forum about this:

14

CPUs are a bit like office buildings, and 100% and 0% CPU usage are a bit like saying the building is Occupied or Unoccupied. It doesn’t tell you much about the operations within.

CPUs have many internal components, which may or may not be in use while it is occupied. A good example of this is a program that gets stuck in an infinite loop will stall and consume 100% CPU, but not make much heat because it’s like only using the lobby of the building. Doing just one thing.

On the other hand, when the CPU fires up AVX-512, the CPU has to downclock aggressively to have enough power for that workload. It’s like if you crammed thousands of people into two floors - you’d have to be mindful of the HVAC needs of those floors, at the expense of other tasks/floors, or everyone there would overheat.

Because of this, no two “100% use” workloads are identical. I am reminded of a graphics driver update a few years back that nVidia put out, which improved one game’s FPS by a whopping 40%. In the immediate aftermath, the forums were plagued with users complaining of crash issues. The drivers did such a better job keeping the components of the GPU utilized, it resulted in too high of temperatures and too high of power draw, causing those crashes. Some power supplies and manufacturer GPU cooling solutions were not up to the task. The solution ended up being underclocking the GPUs with MSI Afterburner.

You think you do a good job by optimizing your software to more fully use your office building, and suddenly nobody can run it because of those crashes! Tsk tsk!

I would start to get to the bottom of it by graphing RAM, CPU and GPU usage. Use programs like RAMMap and MSI Afterburner. Dump the data that you can to a file. I have seen it where the Task Manager thinks RAM is 60% in use, but RAMMap adds up to 97% in use, so more accurate accounting of memory could reveal a point of failure. My main workstation has 128GB of RAM, for good measure.

Graphing CPU, GPU usage, temperatures, boost levels and voltages could also reveal something. A lot of computers will shut down if components exceed 90-100C, so keep that in mind.

Good luck with the detective work ahead of you!

1 Like
15

I was searching for a week and I am so tired. I reinstall Windows (twice) and drivers. Prevent Windows Update. Tried all versions since 1.5.3. Tried on another PC. It a No-Go. Even support team cannot help.

Then, I read an article dated February 2020 comparing video enhancement programs. Surprised, that one of the cons is video size limited to <10 GB. It must have been changed since then because clearly I remember they added support for .mov (ProRes) HQ support on 1.8.1 if I remember correctly. Which means this type of file size can reach hundreds of GB’s easily. And I was so pleased because, I was thinking upscale the old sci-fi series/movies (eg. Babylon 5 movies).

The video I was trying is mxf (ProRes) file which I cleaned up with Premiere Pro neat video plugin. Now I was trying to upscale to mov. (Prores) again so I can do the color corrections on final.

When I read the article it flashes on my mind that this file which I was trying is around 64 GB. Could be the problem? They should have been lifted the 10 GB limit, right?

No. They didn’t. I understand this when I tried another mxf file with the size of 2 GB and all the settings are highest. VEAI handle and finished upscaling in 20 mins.

This is unbelievable and ridicilous. And I don’t understand added support for mxf and mov (ProRes) file and not remove file size limit.

16

It’s really almost always a PSU matter. Tl;dr: you need a quality one – wattage alone is not enough.

1 Like
17

LOL, happened to me too today. Weird. Did a preview of a movie, bam, spontaneous computer reset. That’s no good. I don’t even think VEAI was rendering already (just loading the model). I have an awesome PSU Be-Quiet Dark Power 12 850W; but, like I said, pretty sure there wasn’t even any load on the system. Nothing in Event log either (didn’t expect there to be anything either, tbh).

Let’s just hope this was a fluke.

EDIT: A preliminary inspection points to a bug with nVidia’s new 511.65 drivers, in which (fatal) undervoltage can occur when the card is throttling down too much when idling on your desktop. A suggested remedy has been to set Power Management (Control Panel) to use ‘Max Performance’. Hope that really fixes it.

18

Damn man you need to cool your system better. You don’t want anything running over 80 really. Plus lower temperature = more performance, especially for that graphics card

19

My, I hope those are not your GPU temperatures;

My, I hope those are not your CPU temperatures.

:stuck_out_tongue:

Seriously, you have a very expensive card. Mine (RTX 3080 Ti) will get ca. 70-80C max, as it should. 100C is insane. For your CPU as well. Anything ‘in the low 70ties’, as Intel says, is good. My own i9 12900K gets a bit hotter with all P-cores burning at 100% sustained load (ca 80C), but not 100C, dude. That’s just too much.

20

that post is already a year old,

and i already run my system below 70° because of way better cooling by additional highend fans

1 Like
21

hello guys again just to update the situation with new info the problem is still exist specially on VEAI but this could be something related to windows 10/nvidia but since topaz apps works on GPU could be something related to win10/nvidiaGPU/drivers issue it happened on 3 games (sometimes different driver version) but VERY rare like in last 8 months 3 times computer crashed and restarted by three games but noticed its when i press winKEY+ALT+D but on VEAI it happenes without pressing anything even while leaving computer AFK and leaving VEAI processing just wanted to say because this could be something from windows 10 and related to drivers or something i hope if someone using VEAI on mac can confirm if they even encounter same problem or not if it never happened on mac users then its differently win10/NVIDIA issue and the games that crashed are god of war / snow runner / cyberpunk each game one time so far

22

guys anyway just a heads up those shutting things were from 1.9 - 2.4 versions of VEAI and i didn’t try VEAI for a long time since it made my motherboard somehow lost access to the bios during a shutdown when i was trying to use it but still computer works normal (just i can format or use a usb it make it stuck on start up) yesterday tried version 2.6.3 idk but since yesterday i tried 7 videos all processed without a shutdown and btw i used to run VEAI solo when i had shutdowns but after i saw yesterday there is no shutdown i tested it even while using skype and browser with +40 tabs open for hours today all good, what ever they did ty i hope it stays like this and its not some kind of good luck or something lol

23

Hello Tangerine

I have the same problem with a hard reboot randomly always after the launch of the enhancing process.
It is a hard shut down like using the power button, as if a hardware protection mechanism was triggered.
I have no blue screen and no messages in the windows event log except for the unexpected shut down.

Temperatures stay <60C for the CPU and <70 for the GPU.
No other programs, even FS2020 at ULTRA setting triggers such situation.

I will check the power drawn from the power supply.

Core i7 10700K (stock, not OC)
6900 XT (stock, not OC)
Memory 32G Dual channel supported by ASUS on.a Z490E
Using a AX760 PSU

24

hello bro sorry to tell you i found no solution but i am currently using 2.6.3 and i didn’t update now is rare for my computer to shut down during processing with VEAI so far computer shut down like 3 times since 25 feb somehow the devs changed something that works fine for me ,mostly without intention i think it has something between (window ,GPUs and PSUs) this same problem i noticed it can happen RARELY on some games/softwares with heavy usage of GPU ,currently i am still using NVIDIA GeForce GTX 1070 Ti on win 10 previous VEAI versions used to shut down my computer like 99% rarely when i was able to complete a video but now with 2.6.3 seems all fine made some videos of old footages processing them while browsing/youtube or watching tv shows/movies made them and only 3 times had the shut down on one video maybe because it was low res.

btw i use max memory processing for the GPU since 25 but before 2.6.3 version sometimes was VEAI able to work if i lower it to half memory for 720p videos or lowering it for more for videos below 720p and for 1080p to 4K i make it max sometimes it did work but can’t prove it because sometimes makes computer crash even if i used max memory on 1080p videos or low memory on 720p,

probably if you still have these shutdowns wait for next versions if you using 2.6.4 try go back to 2.6.3 and test but i still didn’t update since 2.6.3 worked fine i stayed with it and i didn’t use VEAI for like last month but tested it just for this reply to make sure its still working seems working fine while writing without any shutdown

25

Hi there again tangieran

So, after a lot of reading and a lot of testing this is what i figured:

  1. 2.6.4 generated much more hard reboots than 2.6.3 and 2.6.2
  2. CPU usage never exceeded 80%, memory usage stayed at 20% and temps never went past 70C even for the 6900XT
  3. The hard reboots looked like a power surge protection which occurred absolutely randomly
    4)…which is surprising because the HW monitor i use indicates that the system never seems to use more than 350W which means the AX760 should have be sufficient.
  4. The original system is rock solid even when running Prime 95 in stress mode for hours on end

But…replacing the AX760 by a HX1000 solved the problem…

So, my conclusion is that VEAI induces a power surge that triggers a protection from the MB or the power supply. This problem is specific to VEAI and is not induced by any other programs i use.

Hope this helps someone…

2 Likes
26

thanks for the info and good luck

28

It’s really almost always the power supply. It never happens to me (RTX 3080 Ti + 12900K).

I did get a 1-time spontaneous reset on start, but that was a documented driver issue (where nVidia erroneously undervolted the GPU under conditions).