Affinity Photo Frequent BSODs

[v_kyr]

15 minutes ago, R C-R said:

Just keep in mind that as has been mentioned, apps can crash but they cannot directly cause a BSOD (AKA a fatal system error or system crash).

Of cause they can result to and initiate BSODs, so you better don't repeat that statement in this general misleading way!

A bad (or intentionally) way coded app accessing common drivers (drivers are just software too, like GPU drivers etc.) can yield to a BSOD. Games that stress the hardware memory & GPU drivers to much can yield to BSODs. Computer virus or another type of malware can cause a blue screen crash. Manipulating registry entries the wrong (or intentionally) way can yield to BSODs.

You can also manually trigger a BSOD on demand for certain purposes, like checking the behavior of your developed software in such cases etc.

See just related some few of the many related references here:

• Malwarebytes
• NotMyFault  <- an app/program for such purposes
• How to manually trigger a BSOD
• How to trigger a Win 10 BSOD on demand
• ... and so on ...

[R C-R]

19 minutes ago, v_kyr said:

Of cause they can result to and initiate BSODs, so you better don't repeat that statement in this general misleading way!

What I said was they cannot directly cause a BSOD, so for example it is the driver that causes the BSOD, not the Affinity app. IOW, the Affinity app calls the driver which in turn causes the BSOD, so the app is the indirect cause.

Do you disagree?

[v_kyr]

4 minutes ago, R C-R said:

What I said was they cannot directly cause a BSOD, so for example it is the driver that causes the BSOD, not the Affinity app. IOW, the Affinity app calls the driver which in turn causes the BSOD, so the app is the indirect cause.

Do you disagree?

Of cause I disagree, since software always runs on hardware and as you should know better, a piece of software (an app/program) can be the root cause for a BSOD. So it doen't matter if you write "indirectly" or not, there are plenty of examples how a software can do so, either intentionally or due to programming faults. And since in such cases the software is the cause and origin there is no indirectly, as the machine (the hardware) woudln't crash otherwise when running a well programmed/ behaving software.

[Old Bruce]

9 minutes ago, v_kyr said:

Of cause I disagree, since software always runs on hardware and as you should know better, a piece of software (an app/program) can be the root cause for a BSOD. So it doen't matter if you write "indirectly" or not, there are plenty of examples how a software can do so, either intentionally or due to programming faults. And since in such cases the software is the cause and origin there is no indirectly, as the machine (the hardware) woudln't crash otherwise when running a well programmed/ behaving software.

So what can the programmers at Affinity do to correct poorly written drivers?

[walt.farrell]

12 minutes ago, v_kyr said:

So it doen't matter if you write "indirectly" or not, there are plenty of examples how a software can do so, either intentionally or due to programming faults. And since in such cases the software is the cause and origin there is no indirectly, as the machine (the hardware) woudln't crash otherwise when running a well programmed/ behaving software.

Only drivers or something else running with system privileges can cause a BSOD.

Yes, other software (such as one of the Affinity apps) can call the driver, and then a BSOD can result because of something the driver did. But at that point, one of the following has happened:

1. The application made a correct call to the driver and passed good data, and the driver passed the data to the system or to the hardware, and the hardware failed.
2. Or, the application made a correct call to the driver and passed good data, and the driver has an error and failed.
3. Or, the application made an incorrect call to the driver, and the driver failed to validate the data that it was passed, and the driver failed.

Only #3 might be viewed as a problem with the application. But even there, the driver is supposed to validate its data, because it runs with higher privileges and therefore cannot trust anything that the application passes to it. So, in case #3 yes, there is an error in the application, but there would not be a BSOD without a more serious error in the driver.

What should happen in #3 is that the driver properly validates the data it has been given, and then the driver crashes the application, not the system. Then you have an application failure, and a clear culprit and something obvious to point the finger at.

[R C-R]

1 hour ago, walt.farrell said:

Only drivers or something else running with system privileges can cause a BSOD.

Which means that the only way for an Affinity app to directly cause a BSOD is if it somehow could be running with system level privileges....

[v_kyr]

1 hour ago, Old Bruce said:

So what can the programmers at Affinity do to correct poorly written drivers?

Together with the user base collecting a list of well behaving third party driver versions (thus checking & evaluating driver versions for overall compatibility).

[R C-R]

5 minutes ago, v_kyr said:

Together with the user base collecting a list of well behaving third party driver versions (thus checking & evaluating driver versions for overall compatibility).

About how many hardware/driver combinations do think that might end up being? Considering how many different GPU's can be installed in Windows  systems & all the GPU & other driver updates that are released at different times by different vendors, there must be a very large number of combos that would need to be evaluated to eliminate the not fully compatible ones.

[walt.farrell]

12 minutes ago, R C-R said:

there must be a very large number of combos that would need to be evaluated to eliminate the not fully compatible ones.

True.

And we know that Drivers can be the trouble point, because we have seen cases here where (with no application changes) updating to a newer Driver can fix a problem, but also cases where updating to a newer Driver can cause a problem. The answer to a user's issue isn't always to install a newer driver. Sometimes it's to install an older one.

[v_kyr]

1 hour ago, walt.farrell said:

Yes, other software (such as one of the Affinity apps) can call the driver, and then a BSOD can result because of something the driver did. But at that point, one of the following has happened:

1. The application made a correct call to the driver and passed good data, and the driver passed the data to the system or to the hardware, and the hardware failed.
2. Or, the application made a correct call to the driver and passed good data, and the driver has an error and failed.
3. Or, the application made an incorrect call to the driver, and the driver failed to validate the data that it was passed, and the driver failed.

Only #3 might be viewed as a problem with the application. But even there, the driver is supposed to validate its data, because it runs with higher privileges and therefore cannot trust anything that the application passes to it. So, in case #3 yes, there is an error in the application, but there would not be a BSOD without a more serious error in the driver.

What should happen in #3 is that the driver properly validates the data it has been given, and then the driver crashes the application, not the system. Then you have an application failure, and a clear culprit and something obvious to point the finger at.

Let's make first clear that it hasn't always just to be the cause of GPU drivers here, there are also other scenarios possible where some registry damage, or a damaged/compromised csrssr.exe (via sxs.dll), etc. can yield to BSODs then. So accessing drivers are just one possible cause here.

For app crashes (no BSODs) due to GPU driver related issues, the Affinity apps during their installation could precheck/test the used GPU driver in a better manner. If they detect operation wise a problem with that one, immediately setup the apps prefs to don't use the GPU hardware performance acceleration (disable) that, or immediatelly tell that it detected during installation and pretesting that the used GPU driver will not be suitable for OpenCL based hardware acceleration at all.

[v_kyr]

35 minutes ago, R C-R said:

About how many hardware/driver combinations do think that might end up being? Considering how many different GPU's can be installed in Windows  systems & all the GPU & other driver updates that are released at different times by different vendors, there must be a very large number of combos that would need to be evaluated to eliminate the not fully compatible ones.

Of course, that's why I said "together with the user base", that would be then similar like collectiing those Benchmark results, aka a list of working GPU + GPU driver version list. - BTW, that benchmark could be enhanced to not only show the used GPU, but maybe also indicate the used driver version and some additional app related infos.

[R C-R]

24 minutes ago, v_kyr said:

Of course, that's why I said "together with the user base"...

Even with their input the list would need constant updating as new drivers for existing GPU's became available, some of which, as I understand it, might be installed without the user being aware that they were. Besides, not all of the Affinity user base regularly if ever posts to or follows these forums.

It's a good idea in theory. I'm just not sure how well it would work out in practice.

[v_kyr]

41 minutes ago, R C-R said:

Besides, not all of the Affinity user base regularly if ever posts to or follows these forums.  ... It's a good idea in theory. I'm just not sure how well it would work out in practice.

Instead of only the forum it would also be possible to place such functionality, collecting the GPU + GPU driver version and showup of supported GPU driver versions so far, also into the apps itself. So a panel under the help menu which friendly asks and allows users to submit (if they want to do so) their GPU driver version for this sort of "...known to work GPU + GPU driver versions list collection). The same way an dynamically updated list (taken from one of the Serif servers) can be shown up on demand, accessable also there from the help menu. -- All in all in a similar accessable fashion as the welcome screen or support & tutorials menu entries etc.

[R C-R]

15 minutes ago, v_kyr said:

So a panel under the help menu which friendly asks and allows users to submit (if they want to do so) their GPU driver version for this sort of "...known to work GPU + GPU driver versions list collection).

Ha! As I am sure you are aware, lots of users ignore the help menu completely, & lots of them do not even really understand what a driver is or its relationship to a GPU.

For Windows users, I think what makes the most sense & is the only practical way to go is for the default not to use hardware acceleration, so it becomes an 'opt in' option where a warning message about possible problems could be displayed.

[walt.farrell]

3 minutes ago, R C-R said:

For Windows users, I think what makes the most sense & is the only practical way to go is for the default not to use hardware acceleration, so it becomes an 'opt in' option where a warning message about possible problems could be displayed.

Unfortunately, I think I must agree with that. 

[Richard H]

It seems my BSOD post has sparked a lively discussion resulting in a lot of thoughtful input. Even with my limited knowledge, I can see what a complex issue this is with no easy solutions . Meanwhile back at the ranch, attached is my BSOD report from today. I also include an image of what I was doing in AP, putting together a presentation of how to identify old Art Nouveau tiles, I managed to do the one slide before the crash. Because of the deadline, nothing for it but to go back to Photoshop, I have just completed fifteen more slides with no problem. 

[R C-R]

15 minutes ago, Richard H said:

Meanwhile back at the ranch, attached is my BSOD report from today.

Note that in the analysis section it says this is likely to be caused by a hardware issue but it could be caused by a misbehaving driver. There is also mention of a possible overheating issue, possibly caused by a bug in a driver.

[v_kyr]

See:

• How to Fix intelppm.sys error Windows 10 (english)
• How to fix error 0xc00002e3 windows 10
• Regular freezing and BSOD on Windows 10 - intelppm.sys seems to be the cause
• ... and so on, do a Google seearch after intelppm.sys ...

The first referenced article from above in german instead:

• So beheben Sie den intelppm.sys-Fehler Windows 10 (deutsch)

[Richard H]

Hi v_kyr,

 Thanks for the links, I speak German fluently so was able to make full use of the information. All test results turned out negative, that’s to say nothing is corrupted or malfunctioning on my system. I have spent far too much time trying to sort the problem out, and short of taking the PC back to the shop for a hardware check, I don’t know what else to do, besides, having the components checked because of one app is a bit extreme. I am going to mothball AP and AD – if updates are released, I will give them a try. In my book the apps get an “A” for what they can do, but a “D” for functionality. I need reliability, and though I don’t want to spend the money, it’s back to Adobe full time for me.

[walt.farrell]

44 minutes ago, Richard H said:

and short of taking the PC back to the shop for a hardware check, I don’t know what else to do, besides, having the components checked because of one app is a bit extreme.

I have lost track of the number of times I have experienced, both on a PC and on a mainframe during my 40 years of experience there, situations where a hardware error seemed to affect only 1 application. The error still existed, and when it was fixed, that application and several others that no one had realized had issues started working better.

So I wouldn't say it's extreme, at all. But I'm not the one who will have to pay for that service call, so it is, of course, your choice

On the other hand, sometimes it's not hardware, as we've mentioned, but drivers. Or issues with the Registry. There have also been several reports here of BSODs and crashes that were resolved by reinstalling (refreshing) Windows, because it tends to get a bit grimy over time which can frequently cause weird problems.

[v_kyr]

For such BSODs the real cause of the origin (the error) can be overall difficult to detect and encircle.

• Wie behebt man CLOCK_WATCHDOG_TIMEOUT Blue Screen (0x00000101)

• BSOD Error code 101

[bobbles1]

Well, this is still happening. I'm yet another user with perfectly up to date drivers, previously running other graphics software that shall remain nameless without a hitch, and trying to make the switch to Affinity. But here I am, in my trial period, and I keep getting blue screen of deathed, only while using Affinity Photo, and always in the middle of something simple, like moving a layer a little bit! Direct X is up to date. Graphics drivers are up to date. It is an older machine with an older graphics card - but if this is the problem then Affinity should be controlling for it, maybe even controlling for it by not letting the software run, but certainly not just blue-screening. Very frustrating.

Edited August 24, 2023 by bobbles1
Accidentally attibuted my post to a previous user because it went in a Reply box.

[Dan C]

Welcome to the Affinity Forums @bobbles1 & I'm sorry to hear you're having trouble.

As previously confirmed in this thread:

On 6/6/2022 at 1:55 PM, Dan C said:

On 6/2/2022 at 3:01 PM, walt.farrell said:

However, the Affinity applications cannot directly cause BSOD errors, as they run with application privileges, and BSODs indicate that something running with elevated (System) privileges has had a fault. When you get a BSOD it generally either means you have a hardware issue, or more likely (with the Affinity applications) a GPU Driver issue.

The Affinity applications aren't able to directly cause a BSOD error - however it's certainly possible the app is using hardware/software that other apps aren't using, which triggers the issue within Windows.

I'd recommend also using Windows Reliability Monitor as this may contain more information as to the cause of this, however the error messages/event viewer screenshots provided do seem to point to a potential hardware issue.

It may be the internal iGPU within your Intel CPU causing this issue, so please make sure these drivers are also up to date, alongside your Nvidia drivers. Equally, please check for any BIOS updates from the manufacturer, as these can also reduce BSOD errors.

I hope this helps!