Apex shutting off PC

Yeah that's what I thought, haven't done any manual defragging for SSDs.

For "Optimising" (which I think means defragging) from the Windows settings though they're on defaults, should I just turn it off for the C drive and see if it helps?

Was also thinking to try not changing the temp limit on Afterburner launch if it's more to do with a driver cleanup after an update.

@Darkness93_yt 

---

Darkness9t3 wrote:

Yeah that's what I thought, haven't done any manual defragging for SSDs.

For "Optimising" (which I think means defragging) from the Windows settings though they're on defaults, should I just turn it off for the C drive and see if it helps?

Was also thinking to try not changing the temp limit on Afterburner launch if it's more to do with a driver cleanup after an update.

---

No, Windows is detecting a SSD and will handle it accordantly, meaning it will do what is right for an SSD.

I said in my last post the SSD don't fragment, but that is not entirely true, even on a SSD it can happen that there are so many fragments on the drive that the Metadata, the data that tells the OS where a specific file is, can hit the limit and that would lead to I/O errors from the SSD.
So let the Optimising on.

Thanks for the info, will leave all unchanged then 🙂

So back to the original issue; I did some testing over this week and didn't change the temp limit at all.

Game still ran fine, which leads me to suspect it's something about the Driver Updates, and when the shaders are rebuilt for the GPU after one, sometimes it will just fail on a few games soon after but then stabalise itself.

There's currently another pending version (493.13, I'm on the previous one; 472.12, not sure why there's such a huge jump in version numbers though) but I'm skeptical to install it because I know Apex will crash out again a few times.

Any tricks with the shader cache I could try maybe? It's going to be something about the shader cache or the driver update, or maybe I should always stay one version behind (manually download it instead of downloading the latest with Experience)?

@Darkness93_yt 

The big jump in version number is probably due to NVIDIA ending support for Kepler GPUs.

For the first time the new driver let you set the shader cache size manually.
But I really doubt that a shader cache problem could lead to a complete shut down of the system.    

@holger1405 @Darkness93_yt I don't understand how not updating drivers is a good idea, as newer drivers usually synergise with the games better and it's always recommended to have everything up to date due to stability?

Yeah normally I'm all for keeping it on the latest driver, but I just have a feeling it'll make Apex crash again; it's always been in line with a GPU update it seems.

@holger1405 I doubt it also but it's a really weird issue, just my bad luck 😛 I'll do the update then I guess and report back after some more gameplays.

Question is; should I use my normal approach of DDU in Safe Mode or just go ahead with an Express or Clean Install? Usually I've used DDU, because when I go Express it's failed anyway. The Clean Install I've done every now and then but it still fails, so I'm not sure which is the most fail-safe approach.

@damsonwhufndthis 

There is a old saying:  Never Change a Running System!
That never counts for security updates!  😉 

Maintaining the System and holding it up to date is of course important, but I also see people "updating" themself into problems all the time.
As a general rule about updating, I would not let my graphic driver get 6 mounts old, but I am still at 471.68, I have no problems whatsoever with the games I play nor with any other task, so why should I update and risk to introduce one?

I also postpone Window updates for a few days to see if there is a problem with them, (That also don't count for important security updates.) and install them if I see that they are fine.

Then, I will update to Windows 11 in the near future, before that I will bring my system up to date, and not only the graphic driver.

The performance gain of a new driver (If any) is in most case absolutely negligible, so the qestion I ask myself is if I need this driver or this update right now? Does this driver has a feature I want?
I would at least suggest to wait a few days and check the NVIDIA or AMD forums to see if there are some problems with a new driver.
But as said, it is also not good the get a graphic driver totally outdated.

@Darkness93_yt 
DDU is a really great tool, but it also a tool that cleans the registry.
Anytime you clean the registry there is a chance that something goes wrong, and wile DDU is a very reliable tool, that risk still exist.
I personally only use DDU if there is problem.

You should shut down about any background program that is running, including the virus scanner and especially tools like Afterburner before installing a graphic driver, but I would suggest to use the normal clean install variant in the driver.

Well so far so good!

After the clean install update for the graphics driver I've not had a crash again over the week, seems pretty stable on Apex now! Temps for the GPU are still good because the custom fan curve is there, but I leave the temp limit unchanged as there's no need for that it seems plus there's no extra power draw.

I guess the current driver has more support for certain GPUs, but a clean install seemed the way to go.

Thanks so much for your help guys! I'll pass it back onto C-Sheldon88 on the other thread (https://answers.ea.com/t5/General-Discussion/apex-legends-restarts-my-pc/td-p/10080004/page/3) to see if it helps him too (would be good to confirm as he's on a different GPU).

It's been great really digging into what the issue could be, hope this thread helps other users with the same issue!

Guess I spoke to soon, happened again tonight.

There was a pending driver update though (don't know if that has anything to do with it), so I'm trying a clean install of that.

I'll change the temp limit again too to be safe, as that's what made it more stable in the past; it must somehow spike to the stock temp still even with the fan curve on.

@Darkness93_yt at this stage, if I was you, I'd be totally bald from pulling my hair out! It really is a bizarre issue. 

All that remains is to put the computer in a fridge or next to an ac and run apex on it. That will tell you definitively if somthing is broken/ill fit or it's a settings/software issue. 🙂

I have had this issue recently as well. And I'm running on a 3060 with a 650 Seasonic 80+ Gold PSU, so I am not sure this is a hardware issue. As of late I haven't had many crashes, so I actually kinda think this is an Apex issue given that it changes with time.

Also, I've stress tested my computer at build with Furmark and my machine learning use case. No issues there. Even one of the CPU stress test programs that I used (something from CPUID) running on the SSE instruction set caused my CPU to hit its temperature limit and still nothing bad happened (that benchmark just happened to be very good at cooking the CPU; everything else keeps it in the low 70s so I'm not sure how CPUID did it)

Hey @Darkness93_yt what are your PCs specs (including the names and also give me the memory speed)? May have an answer for you that I found Googling.

@Darkness93_yt Sorry for such a late response. I found something (the thread at https://forums.overclockers.com.au/threads/new-build-b550m-shutting-down.1292677/page-2 and other reference(s)) that suggested that turning off XMP in the BIOS solves this issue on, at least (to the best of my recollection) B550 motherboards. I did that and it worked. However, in that thread, they also mention on page two of that thread that there's a new BIOS version that solves the issue, so XMP can be turned back on once it is updated. This ended up working for me.

In theory you can update without doing the whole XMP shutoff thing, but if the computer shuts off in the middle of a BIOS update, you may brick your motherboard.

Yeah I updated my BIOS not long ago, I'll double check my XRP setting though as it happened again tonight but I'm out of the habit of checking the Temp Limit setting on Afterburner even though the Fan Curve is working.

A project I need to get onto really is putting proper thermal pads on the card, or maybe I should just start browsing for an RTX 3090ti or so where they've sorted the whole overheating issues.

It's not a great help in the summer where I have to use Afterburner and fans / my portable air con on full blast to combat it all! 😛

Had a good run though so far this year without it overheating, so I'm sure the Afterburner stuff is helping.

So interesting flurry of tests I've had to do over the past few days due to lots more crashes, and I even managed to grab a pic at the point of crash.

Interesting to note that I did just apply a Windows Update before seeing all the crashes, which probably disrupted something about my IGPU state or so.

I think I'm homing in on the main issue though as it's been more stable after verifying my Integrated GPU status among some other things, I'll walk you through it:

Apex started doing the weird thing again where the graphics would go all fuzzy and my PC would restart a few seconds later, costing me valuable match time and RP (if I'm not able to rejoin in time).

I had a go at disabling XRP for my RAM which didn't help at all and actually made it more frequent, so I switched it back on again.

I then updated my BIOS to the latest version, just in case something weird was going on there.

I reinstalled Apex, and updated once the new Season came out, but it was still crashing (about the time I was able to grab the pic of it).

With the crashes still happening, I looked again in Apex's Crashpad folder (last crash on 10th August) and thought harder about the vague error message it gave;

ERROR_CODE: (NTSTATUS) 0xc0000005 - The instruction at 0x%p referenced memory at 0x%p. The memory could not be %s.

 Apart from it being a general memory reference issue, the key here was the NTSTATUS part, relating to some function of Intel's HD Integrated Graphics Driver stuff after some quick Googling:

https://forums.techguy.org/threads/error_code-ntstatus-0xc0000005.1262544/

I went ahead and downloaded the latest Intel HD Graphics Driver, and was able to install only after going back into my BIOS and specifically Enabling Intel HD Graphics (by default it was set to "Auto").

The game has been very stable since and I've been able to enjoy some great Kings Canyon matches on Ranked for a good few hours each time.

My feeling is that Apex is still somehow trying to send instructions to all graphics cards detected on the system, one of which being the Integrated graphics (even if there's a physical GPU there, I'm still using the RTX 3090). If IGPU is Disabled or on Auto in the BIOS, I think it doesn't manage to completely switch over in time, which is causing the crashes.

Hope the Apex Devs are able to sort out all these crashes in general though cause there's loads of forum posts about it, but hopefully I've touched on something here that will help them verify their GPU code.

Again, this only happens with Apex Legends. I can run any other game fine on my PC, it's just Apex that crashes.