PC Graphics

[andreysolo]

Hello. I bought a RX560 about 1,5 years ago and from the start met driver crashes, without any dependence of the usage of card - she can crash on the youtube or not very "heavy" games from 2008-2010 years as Mount and Bladeю Also I notice freezes of the image, on the youtube or when I watch video at the Itunes player
My current configuration:

PSU: Zalman AC600 600W

Motherboard: Foxconn P55MX

CPU: Intel i5-650 3,2 Ghz

RAM: Patriot Signature Line 2x4 GB 1333Mhz

GPU MSI RX560 AERO ITX

I used Furmark and MSI Afterburner/HWinfo earlier. Inb the furmark test GPU temperature wasn't above 63 deg C, HW Info says that GPU memory temperature wasn't above 50-55 deg С
But using graphs on the HW Info I see, that amperage on the card's core jumps to around 15-18 A and for some cases even to 29-36 A
Can this be mini - short circuit from my old PSU or another problem from this "direction"?

Memory Usage in moment of crash 

Jumps of Amperage 

[FileNotFound__]

So the high amperege isn't a problem, the core uses about 1.2v, using the formula P=UxI you get about 34.9W power consuption at 29A.

So the random crashes and BOSDs pull more towards a fulty harddrive.

Are there any arteffects or something that makes you suspitious of your graphics card other than the high amperege.

[andreysolo]

Hmm, artefacts are not very frequent with this issue, for 1,5 years I see them around 4-5 times. Almost always crash looks like blackscreen with attempts to restore drivers (fast change of the image on display, freezes of the image)
Also in few cases of crash while I'm playing in the Mass Effect 2, I heard soud like "drrr" i. e. my fan propeller is rattling, chattering. But, maybe, can this be a HDD trouble?
Also while I listen music or watch video I hear crackling in my headphones, but I don't sure that this is consequence of this problem and not of bad organisation to my forward panel and 3,5 mm jacks
About HDD.. Yes, I also think about this, because start to see at the process monitpr. that disk almost always has a 100% of active time and actively uses pagefile.sys, but I think, that doubling of RAM can help, but I also has these problems
Also my PC begins not to go out from the sleeping mode correctly - dispaly's backlight is there, but no image.

I attach the CrystalDisk info about my disk, can you say, is this all right there?
Also I use chkdsk from Windows some times - it doesn't find any mistakes

CrystalDisk info 

[FileNotFound__]

The arteffect you sent looks like a GPU issue, and GPU fan rattling defenetly means it isn't getting propper cooling. But before going straight for the gpu, maybe you should do a clean driver instalation.

After that you should install GPU-Z from techpowerup to check how hot the gpu is getting and the behaviour of the gpu when the problem occurs.

To test the harddrive I would reccomend you use HDD sentinel trial, it will straight up tell you the condition of the drive and amount of bad/unreadabale sectors.

[andreysolo]

Sorry, but another user sent you a screen of artiffects, this isn't my situation. I use GPU-Z and MSI Afterburner, their graphs says that temperature of GPU at the calm state (without active usage) is about 28-30 deg C, for the usage - not above 60 deg.
Behavior on the issue... I can show you a screen from Afterburner graph which I managed to take after crash:

I see there a) big peak of thetemperature with strange indicator as 511 (!) deg С (this is a mistake, I think), and b) fast falling of VRAM memory usage to 0 MB or around thisGraph of MSI Afterburner, downing of the usage of memory, peak of temperature of gpu

Sorry for russian interface, I hope, that graph will be clear without translation

[FileNotFound__]

Seems normal, exept the spike but that spike might actually be your problem. The spike might be caused by the communication issue, try to lightly shake your system while it is running if that triggers the fault, you got a bad connection somewhere.

This is a legit method that we use sometimes when we have an intermittent problem btw.

Also try installing the mentioned HDD sentinel to see your harddrive stasus.

I should have asked this before, but is there a specific action that triggers the fault or is it just completely random?

[andreysolo]

"Shake a system" - do you mean physically shake a system unit?
About Sentinel - trial version says, that my HDD - both internal and external - are "excellent", their health this program ocunts as 99-100% (I make a screen)

HDD Sentinel situation
About specific action... I can't say, that this is something specific as cause of the crash.
From one hand by experimental way I recognised some games, which frequently stay me in front of this issue: Mass Effect 2, Mount and Blade and her mods, Battle Brothers, Halo: Anniversary Edition (with old graphic). So, you can see, that diapason of "heaviness" for these games is wide - from old AAA projects to the not very demanding indi- or semi-indi projects, which demand as max graphic card even weaker than RX560.
And from the other hand I don't have these issues on the other "heavy" projects like Assasin's Creed Revelations. Yes, this is not the very good example, but I think, that this is something equivalent to the ME2 and far more heavier than Mount and Blade or Battle Brothers or other indi-games.
(Also about gaming situation... I see the problem with loading of texture at the Witcher 2 Assasin of Kings - they are not load or load very slowly, but this crash doesn't occur).
I can't understand rightly situations, which lead to these crashes, but I notice that frequently this occurs when I click by mouse, even when I exiting from the game
Now I also see these crushes when I watch m4v file- frequentlier when I use Itunes, rarelier - when I use Media Player Classic
From time to time this occurs while I watching youtube videos at 720p. also I have microfreezes of the image on youtube, especially when I jumps to Youtube from othe browser's tab

P. S. Also about graphs from my firts message - all in all it seems strange for me, that amperage of Core has a spike from 0-3 A to 15 A...

[FileNotFound__]

Yes, i meant lightly shake an entire system, just enough to indicate if there are any  bad/dirty connections but not hard enough to cause damage.

Nice that your harddrive is in good shape, but that means we are still far from the issue.

So, the computer crashes on all games no matter the intensity? The fact that your system crashes  while video playback, does indicate there has to be video decoding type of load to trigger the crash. Does your system crash while just idling on the desktop?

About the current spikes, I doubt thats a problem. Every spike in GPU utilisation will cause a spike in current draw, as far as I know thats a completely normal behaviour.

So the next step you should do is the "earthquake test" or at least that is what I call violently shaking someones system, and if that doesn't trigger any issues, next step would be A clean driver install. You do that by installing DDU (amd version), booting into safe mode, running the app and resseting the system. Then you just install newest amd software from amd's website and you should be good to go. Booting into safe mode is necesary as in safe mode, the computer doesn't start any drivers so you can easily delete them without the computer fighting you.

[andreysolo]

Ok, I'll try to run the probleming game and shake my system
About drivers...
I should say that I early had AMD HD 5670, which I bought in 2010 (while I managed system), but at 2020 she was "died" with similar sympthomatic (black screens, crash of drivers). While I waited my RX560, I used as "erzatz" Nvidia GT 1030.. And for first time I make driver installation for AMD without deleting Nvidia driver... But when crashes began, I make some "clean uninstall" both by Display Driver Uninstall and AMD Cleanup Utility, tried to install different versions of drivers, but problem remains

And for Sentinel and HDD... Mm, It seems for me that there is a mistake at the Sentinel account about my system, because my HDD is about 10-11 years old, and I used it... a bit hard

[FileNotFound__]

Harddrive life time is purely based on luck, I repaired a lot of computers I have harddrives ranging from 19 years old that work great and have no signs of stopping, but I lso have some brand new drives that just fail, doesn't matter if you are using it a lot or if it is just running as your secoundary drive full of screenshoots you will never open again. But when a drive starts to fail, it is gone. 

Based on the driver history you said, I got a feeling that your computer is just full of remains of old drivers and that might be causing a problem  what I recomend you to do is:  

1. go to this site https://www.guru3d.com/files-details/display-driver-uninstaller-download.html and install ddu

2. Go to amd support https://www.amd.com/en/support and choose your specific GPU driver and download the installer, do not run it.

3. Disconnect your computer from the internet and boot into safe mode

4. run DDU for both nvidia and amd, even intel if there is an option and make it delete EVERYTHING exept things that say at your own risk

5.restart your computer (ethernet still disconnected) so windows doesn't try toinstall basic ones

6. run driver installer (offline) and let it reset the system

7. If the instalation is succesfull, connect to the ethernet and you are set.

8. If the problem is still there, eather you need to reinstall windows or yor GPU is cooked.

[andreysolo]

Ok, I’ll try your advice, but I already try this few times with different version of drivers from the AMD official site without any success

I’ll try and something went wrong - while I installing driver at the safe mode, process of installing suddenly jumps from ~60 to 80% and then, on 88% install crashes with error 192

[FileNotFound__]

No, you were suppoused to run DDU 3 times in safe mode once for AMD, once for NVIDIA and once for INTEL, after that EXIT safe mode and boot your sytem normally and then install the driver and if it is succesfull, you can connect to the internet.

You can't run driver installer because it will try to activate it, but windows in safe mode will block any drivers from starting, giving you errors.

[andreysolo]

I did yesterday what you adviced me about clean setup and I didn't have this problem, but today, right now, I faced that at the Battle Brothers game, but, instead of a real "black screen"  fortunately, this was "delay in driver detection", as AMD Bug report said me.
On the MSI Afterburner graph I notic ehard pincrease of teh GPU utiilization to the 100% with falling of VRAM utilization to the 60 MB... Without any serious increasing of GPU temperature

Peak of GPU, fallingf of memory

[FileNotFound__]

It is completely fine for the GPU to peak while exiting the app and vram falling down to idle.

I am happy your System no longer crashes, but unfortenually, I never heard of "delay in driver detection" and I am unable to find any info online.

Is that an Issue you can replicate?

Can you give me more information I could work with?

Does it happen with any other game?

[andreysolo]

Sorry, please, if my tone seemed you offend. I simply tried to describe a problem
About replication of issue - sorry, but I can't understand, what can lead to these crashes
For example, about gaming - all can be stable for, let's say, few hours - no freezes and artiffects, good performance, temerature at the RTSS overlay around 35-40 deg C. And then accidentally image is stopped and display starts to go out and return few times.
Early there was a BSOD frequently, now driver restores and I take a message from the AMD Bug Report tool
I also play another game from my "list of problems" - Halo Combat Evolved Anniversary, there also wasn't any of these issues
When I will have this at the next time, I'll try to copy and translate this message (Sorry, I don't think that screenshot with Russian text can help )
P. S. I notice now that after installation of latest version of driver I can't open AMD Settings, because there is no Radeon Software.exe at the "C:\Program Files\AMD\CNext\CNext"path

P. P. S. I had advice from one russian forum, that this can be technical problem with errors in VRAM on the card ( as I understand, there can be bad bytes on the chips of memory)
Can we tell me (if you know), how can I test VRAM for errors?

[FileNotFound__]

You didn't offend me, I just daid i was unable to find any refrences.

Have you installed just the amd driver or the amd software that comes with the driver.

Sadly while theese issues do sound like a bad GPU, there is still a chance of something else might be wrong, and if you do not have another system to test it we have to go on, testing it step by step.

And unfortenually, I will have to inform you that there is no cure for the VRAM errors. I came to this community because I have a GPU with a faulty VRAM as you can see here https://community.amd.com/t5/graphics/r9-270x-memory-diagnostics/m-p/512221#M83239 and noone came with a solution.

[andreysolo]

@FileNotFound__

About description of the problem at the message from the AMD Bug Report Tool, translated:
"A driver response timeout has been detected in your system"

""And unfortunately, I will have to inform you that there is no cure for the FRAME errors. I came to this community because I have a GPU with a faulty VRAM as you can see here https://community.amd.com/t5/graphics/r9-270x-memory-diagnostics/m-p/512221#M83239 and noone came with a solution."

Sorry, I understand, that only one way to cure VRAM is replacement of bad chips. But firstly I had advice to test RAM by Archlinux instrument or Server (especially for AND devices). But, as I understood, that should be another video card/integrated graphic to use this? So that card which is the aim of testing i shouldn't be used?
Sorry, but I don't have system to test... But I try to use multitester to my PSU, because I nearly (two years ago) had almost the same issues with my old card which all in all cooked.

[FileNotFound__]

You can try testing the VRAM with OCCT software

[andreysolo]

Sorry, does this safe enough? If I understand right, ЩССЕ is for stress testing, can this lead to the physical failure of graphiс card?

[FileNotFound__]

Well If your GPU is really at its last legs, stresing the faulty part could end it, but It won't harm a healthy card, in fact after stressing the crap out of my R9 270x while trying to figure out if it is the memory or the core that its the issue, id made it semi usabale, as long as I don't try to change the frequency or use the video enoder it now works fine.