AMD issues.

S

Swinny

Bronze Level Poster
Hi all.

I bought a custom build pc from here a couple months back and alls been great.

However lately when playing FF14 out of the blue my pc screens go black - after about 30 seconds I get the error message

AMD Software detected that a driver timeout has occurred.

I’ve uninstalled and reinstalled the drivers I’ve cleared cache and sometime that will stop it for a day or so but then it just comes back and happens again.

Any ideas how to resolve this please.

Full pc specs below

Case


CORSAIR 5000D AIRFLOW TEMPERED GLASS GAMING CASE - WHITE


Promotional Item


Get a discount code for 20% off select peripherals at Corsair.com


Processor (CPU)


AMD Ryzen 9 7950X3D 16 Core CPU (4.2GHz-5.7GHz/144MB w/3D V-CACHE/AM5)


Motherboard


ASUS® ROG STRIX X670E-A GAMING WIFI (AM5, DDR5, PCIe 5.0, Wi-Fi 6E)


Memory (RAM)


32GB Corsair DOMINATOR TITANIUM DDR5 6400MHz (2 x 16GB) KIT


Graphics Card


24GB ASUS TUF GAMING RADEON™ RX 7900 XTX OC - HDMI, DP - DX® 12


Graphics Card Support Bracket


NONE (BRACKET INCLUDED AS STANDARD ON 4070 Ti / RX 7700 XT AND ABOVE)


1st M.2 SSD Drive


500GB SAMSUNG 980 PRO M.2, PCIe NVMe (up to 6900MB/R, 5000MB/W)


1st M.2 SSD Drive


2TB SAMSUNG 990 PRO M.2, PCIe 4.0 NVMe (up to 7450MB/R, 6900MB/W)


1st Storage Drive


2TB SEAGATE BARRACUDA SATA-III 3.5" HDD, 6GB/s, 7200RPM, 256MB CACHE


Power Supply


CORSAIR 850W RMx SERIES™ MODULAR 80 PLUS® GOLD, ULTRA QUIET


Power Cable


1 x 1.5 Metre UK Power Cable (Kettle Lead, 1.0mm Core)


Processor Cooling


CORSAIR iCUE H150i ELITE CAPELLIX XT RGB High Performance CPU Cooler - WHITE


Thermal Paste


ARCTIC MX-4 EXTREME THERMAL CONDUCTIVITY COMPOUND


Extra Case Fans


5 x Corsair AF120 RGB ELITE PWM Fan + Controller Kit


Sound Card


ONBOARD 6 CHANNEL (5.1) HIGH DEF AUDIO (AS STANDARD)


Network Card


ONBOARD 2.5Gbe LAN PORT


USB/Thunderbolt Options


2 PORT (2 x TYPE A) USB 3.0 PCI-E CARD + STANDARD USB PORTS


Operating System


Windows 11 Home 64 Bit - inc. Single Licence


Operating System Language


United Kingdom - English Language


Windows Recovery Media


Windows 10/11 Multi-Language Recovery Image - Unlimited Downloads from Online Account


Office Software


FREE 30 Day Trial of Microsoft 365® (Operating System Required)


Anti-Virus


NO ANTI-VIRUS SOFTWARE


Browser


Firefox™


Warranty


3 Year Silver Warranty (1 Year Collect & Return, 1 Year Parts, 3 Year Labour)


Delivery


STANDARD INSURED DELIVERY TO UK MAINLAND (MON-FRI)


Build Time


Standard Build - Approximately 3 to 5 working days


Miscellaneous


PCSpecialist 2023 Keyring


Welcome Book


PCSpecialist Welcome Book


Logo Branding


PCSpecialist Logo


Packaging


PCS Corsair Carton Upright


Price: £0.00 including VAT and Delivery


Unique URL to re-configure: https://www.pcspecialist.co.uk/saved-configurations/amd-am5-pc/ggC97rrf4F/
 
Last edited:
Martinr36

Martinr36

MOST VALUED CONTRIBUTOR
Can you please post the full spec from your orders page, also take a read of this

When posting for Technical Support...

Hi all If you're here, it likely means you're having issues which is not good, so we want to help to make it much better as quickly as possible and get you back into full tilt gaming / rendering / production or whatever is your poison. It really helps us if we have the full system...
www.pcspecialist.co.uk www.pcspecialist.co.uk
 
S

Swinny

Bronze Level Poster
Martinr36 said:
Can you please post the full spec from your orders page, also take a read of this

When posting for Technical Support...

Hi all If you're here, it likely means you're having issues which is not good, so we want to help to make it much better as quickly as possible and get you back into full tilt gaming / rendering / production or whatever is your poison. It really helps us if we have the full system...
www.pcspecialist.co.uk www.pcspecialist.co.uk
Click to expand...
Apologies - amended with specs
 
SpyderTracks

SpyderTracks

We love you Ukraine
It's a bug in the game, found this:

‎06-06-2023 09:56 AM

​

Hello,

Sorry if this is a dead post, but I have found something that has worked for my setup in particular. My setup:

CPU: AMD R9 5950X
GPU: XFX Merc 310 7900 XTX
Motherboard: ASUS X570 Plus Wi-Fi
PSU: 1000W EVGA
RAM: 64 GB DDR4 3600
Monitor: Samsung 32 G7 1440p 240Hz
OS: Windows 11 22H2
AMD Driver: 23.5.2
I think there is a problem with the Adrenalin software itself being able to recognize GPUs' clock rates. I had noticed that no matter what game I played my fans seemed to be at 100% before long and then game would crash at some point. Adrenalin would show GPU clock at around 2900MHz which even the boost clock on my GPU model is only supposed to be up to 2615MHz. I think this was causing overheating or something similar. I went into Performance Tuning in Adrenalin and set a default profile. I set the GPU clock to the manufacturer's published max game clock rate (2455MHz) and changed max fan speed to 75%.
I then created a profile for Final Fantasy XIV in which I set the clock rate to a max of 2350 MHZ (-100MHz of max).

I still get the same or better fps in games I have tested so far with max settings:

Control (Max Settings, Max RT): ~85 FPS
FFXIV (1440P, Maximum settings): 120-400 FPS

Cyberpunk (1440, Ultra settings, Max RT, FSR 2.1): ~60 FPS

Your mileage may vary of course, but this has been working for me so far with 0 crashes and GPU junction temp staying in low to mid 80s while under heavy load.
I hope this helps!
1 Like
Reply





Slashic

Slashic
In response to AngelofKaos
Adept II

‎06-07-2023 04:24 AM

​

Yes this seems to be the root cause of the issue.
Most likely the GPU governors not managing correctly the frequencies and voltage of the GPU for very short amount of time maybe, maybe microseconds? Not sure but it's impossible to see with standard monitoring.
Boost behaves differently from games to games too that must be why it's only impacting FFXIV and maybe some other games?
For my model (reference model), creating a profile with a boost at 2700mhz and undervolt at 1130mv fixes this.
This doesn't alter performances at all since it never boosts above 2700mhz in that game anyway, but the governor seems to behave differently and no crashes so far after 2 months.
2 Likes
Reply





jurifrog

jurifrog
In response to AngelofKaos
Journeyman III

‎06-16-2023 01:27 PM

​

I tried this and haven't crashed yet! Thanks for your discovery. My Adrenalin is also showing the GPU clock super high even though my 7900 xtx is only supposed to go up to 2500 MHz, so weird.

Source: https://community.amd.com/t5/graphics-cards/7900-xtx-crashing-in-final-fantasy-xiv/m-p/572374/page/2
 
S

Swinny

Bronze Level Poster
SpyderTracks said:
It's a bug in the game, found this:

‎06-06-2023 09:56 AM

​

Hello,

Sorry if this is a dead post, but I have found something that has worked for my setup in particular. My setup:

CPU: AMD R9 5950X
GPU: XFX Merc 310 7900 XTX
Motherboard: ASUS X570 Plus Wi-Fi
PSU: 1000W EVGA
RAM: 64 GB DDR4 3600
Monitor: Samsung 32 G7 1440p 240Hz
OS: Windows 11 22H2
AMD Driver: 23.5.2
I think there is a problem with the Adrenalin software itself being able to recognize GPUs' clock rates. I had noticed that no matter what game I played my fans seemed to be at 100% before long and then game would crash at some point. Adrenalin would show GPU clock at around 2900MHz which even the boost clock on my GPU model is only supposed to be up to 2615MHz. I think this was causing overheating or something similar. I went into Performance Tuning in Adrenalin and set a default profile. I set the GPU clock to the manufacturer's published max game clock rate (2455MHz) and changed max fan speed to 75%.
I then created a profile for Final Fantasy XIV in which I set the clock rate to a max of 2350 MHZ (-100MHz of max).

I still get the same or better fps in games I have tested so far with max settings:

Control (Max Settings, Max RT): ~85 FPS
FFXIV (1440P, Maximum settings): 120-400 FPS

Cyberpunk (1440, Ultra settings, Max RT, FSR 2.1): ~60 FPS

Your mileage may vary of course, but this has been working for me so far with 0 crashes and GPU junction temp staying in low to mid 80s while under heavy load.
I hope this helps!
1 Like
Reply





Slashic

Slashic
In response to AngelofKaos
Adept II

‎06-07-2023 04:24 AM

​

Yes this seems to be the root cause of the issue.
Most likely the GPU governors not managing correctly the frequencies and voltage of the GPU for very short amount of time maybe, maybe microseconds? Not sure but it's impossible to see with standard monitoring.
Boost behaves differently from games to games too that must be why it's only impacting FFXIV and maybe some other games?
For my model (reference model), creating a profile with a boost at 2700mhz and undervolt at 1130mv fixes this.
This doesn't alter performances at all since it never boosts above 2700mhz in that game anyway, but the governor seems to behave differently and no crashes so far after 2 months.
2 Likes
Reply





jurifrog

jurifrog
In response to AngelofKaos
Journeyman III

‎06-16-2023 01:27 PM

​

I tried this and haven't crashed yet! Thanks for your discovery. My Adrenalin is also showing the GPU clock super high even though my 7900 xtx is only supposed to go up to 2500 MHz, so weird.

Source: https://community.amd.com/t5/graphics-cards/7900-xtx-crashing-in-final-fantasy-xiv/m-p/572374/page/2
Click to expand...
Thank you for getting back to me on this really appreciate it.

I think I’ve managed to follow the instruction and do it as I’m not ofay in this area.

Thanks again
 
S

Swinny

Bronze Level Poster
Hi all,

So a little update.

Today I bought BlackMyth Wukong.

5 minutes into the game and it’s done exactly the same thing?

Really quite frustrating, what can I do?
 
ubuysa

ubuysa

The BSOD Doctor
Can you please download and run the SysnativeBSODCollectionApp and upload the resulting zip file to a cloud service with a link to it here. The SysnativeBSODCollectionApp collects all the troubleshooting data we're likely to need. It DOES NOT collect any personally identifying data. It's used by several highly respected Windows help forums (including this one). I'm a senior BSOD analyst on the Sysnative forum where this tool came from, so I know it to be safe.

You can of course look at what's in the zip file before you upload it, most of the files are txt files. Please don't change or delete anything though. If you want a description of what each file contains you'll find that here.
 
S

Swinny

Bronze Level Poster
Hi
ubuysa said:
Can you please download and run the SysnativeBSODCollectionApp and upload the resulting zip file to a cloud service with a link to it here. The SysnativeBSODCollectionApp collects all the troubleshooting data we're likely to need. It DOES NOT collect any personally identifying data. It's used by several highly respected Windows help forums (including this one). I'm a senior BSOD analyst on the Sysnative forum where this tool came from, so I know it to be safe.

You can of course look at what's in the zip file before you upload it, most of the files are txt files. Please don't change or delete anything though. If you want a description of what each file contains you'll find that here.
Click to expand...
Sorry for the late reply ive been on nights.

I think ive done it right...


Also a couple of reports stated they took to long to generate to were skipped,
 
ubuysa

ubuysa

The BSOD Doctor
From your Application log there are dozens and dozens of messages relating to an app crash that Windows has recovered from, at least one relates to amdfendr.sys, this is a component of the AMD Crash Defender System...
Code: 
Log Name:      Application
Source:        Windows Error Reporting
Date:          23/08/2024 21:54:09
Event ID:      1001
Task Category: None
Level:         Information
Keywords:      
User:          SYSTEM
Computer:      Swinnys
Description:
Fault bucket LKD_0xA1000005_amdfendr!unknown_function, type 0
Event Name: LiveKernelEvent
Response: Not available
Cab Id: fd82c4b8-526c-40f5-9db7-18c8986e2594

Problem signature:
P1: a1000005
P2: 0
P3: 0
P4: 0
P5: 0
P6: 10_0_22631
P7: 0_0
P8: 768_1
P9: 
P10: 

Attached files:
\\?\C:\Windows\LiveKernelReports\AMD_WATCHDOG\AMD_WATCHDOG-20240823-1953.dmp
\\?\C:\Windows\SystemTemp\WER-216919062-0.sysdata.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.dffe47f6-eb00-46b6-9b9c-22b911d80d52.tmp.WERInternalMetadata.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\ReportQueue\Kernel_a1000005_86b5bb5529d530f239d74f6845197146ca9f4057_00000000_cab_1790a091-b266-4c01-bfe1-da713606c9f1\WPR_initiated_DiagTrackMiniLogger_OneTrace_User_Logger_20240808_1_EC_0_inject.etl
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.4d30c19c-70ec-44e9-bd16-e4b68f4b91e0.tmp.etl
\\?\C:\ProgramData\Microsoft\Windows\WER\ReportQueue\Kernel_a1000005_86b5bb5529d530f239d74f6845197146ca9f4057_00000000_cab_1790a091-b266-4c01-bfe1-da713606c9f1\WPR_initiated_DiagTrackMiniLogger_WPR System Collector_inject.etl
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.119649c3-8804-432e-a8e6-c3fdbe82d65a.tmp.etl
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.c3312e9f-3922-4690-a43b-beca745c21bf.tmp.csv
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.c65bfd49-677f-4bdb-925f-a8f7330040bb.tmp.txt
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.f2d707fd-f738-4198-9446-b1a0fa60bd71.tmp.xml
\\?\C:\Windows\SystemTemp\WER.5629a733-22be-487d-9247-d26929fdcecd.tmp.WERDataCollectionStatus.txt

These files may be available here:
\\?\C:\ProgramData\Microsoft\Windows\WER\ReportArchive\Kernel_a1000005_86b5bb5529d530f239d74f6845197146ca9f4057_00000000_cab_1790a091-b266-4c01-bfe1-da713606c9f1

Analysis symbol: 
Rechecking for solution: 0
Report Id: 1790a091-b266-4c01-bfe1-da713606c9f1
Report Status: 268435456
Hashed bucket: 
Cab Guid: 0
If you look in the folder C:\Windows\LiveKernelReports\AMD_WATCHDOG\ you'll see many live kernel dumps related to these crashes, please upload the most recent eight dumps (check the timestamps). You can zip them up first if you like.

There are other curious errors and warnings in both your Application and System logs but these could all be related to an apparently flaky AMD Crash Defender System. I'll know much more when I can analyse those live kernel dumps.
 
ubuysa

ubuysa

The BSOD Doctor
Thanks for those, they do make thinks a lot clearer. The dumps are all identical (as I expected) and all of them have an UNKNOWN_BUGCHECK_CODE (0xA1000005) which means that the kernel detected an error but wasn't able to categorise what type of error it was.

The problem in all the dumps is with amdfendr.sys, the AMD Crash Defender System. This is supposed to recover from hangs in your AMD graphics card and/or driver. In the dumps we see the AMD graphics driver (amdkmdag.sys) called repeatedly, which is normal and expected, but then we see amdfendr.sys called, that indicates that a graphics hang was detected by the AMD Crash Detection System and recovery was attempted. You can see this in an extract of the function call stack (which you read from the bottom up)...
Code: 
ffffb688`e9f2e818  fffff804`c16c34d8 amdfendr+0x534d8
ffffb688`e9f2e820  00000000`0000027f
ffffb688`e9f2e828  00000000`00000000
ffffb688`e9f2e830  00000000`00000000
<empty stack records skipped>
ffffb688`e9f2ec38  fffff804`c16874ba amdfendr+0x174ba
ffffb688`e9f2ec40  00000000`00000000
ffffb688`e9f2ec48  00000000`00000000
ffffb688`e9f2ec50  00000000`00000000
ffffb688`e9f2ec58  00000000`00000000
ffffb688`e9f2ec60  00000000`00000000
ffffb688`e9f2ec68  00000000`00000000
ffffb688`e9f2ec70  ffffb688`e9f2ed40
ffffb688`e9f2ec78  00000000`00000000
ffffb688`e9f2ec80  ffff60f3`cd37c3c0
ffffb688`e9f2ec88  ffff938c`559b08d0
ffffb688`e9f2ec90  00000000`00000000
ffffb688`e9f2ec98  00000000`00000008
ffffb688`e9f2eca0  00000000`00000000
ffffb688`e9f2eca8  00000000`00000006
ffffb688`e9f2ecb0  00000000`00000008
ffffb688`e9f2ecb8  00000000`00000000
ffffb688`e9f2ecc8  fffff804`c1674b33 amdfendr+0x4b33
ffffb688`e9f2ecd0  ffff938c`4b913000
ffffb688`e9f2ecd8  00000000`00000000
ffffb688`e9f2ece0  fffff804`dce293c0 amdkmdag+0x14093c0
ffffb688`e9f2ece8  fffff804`dbefc415 amdkmdag+0x4dc415
You see at the bottom the amdkmdag.sys graphics driver being called a couple of times at different offsets. This is perfectly normal behaviour. Then we see the first amdfendr.sys call and this must be because the AMD Crash Detection System detected (or thought it detected) a graphics hang. Then we see a second call to amdfendr.sys at a different offset, followed by a third call at a different offset again and that's where the function call stack stops, so the error was in that third amdfender.sys function call.

This is confirmed by the dump triage analysis which also points at amdfendr.sys...
Code: 
FAILURE_BUCKET_ID:  LKD_0xA1000005_amdfendr!unknown_function
There you see LKD meaning this is a live kernel dump, the UNKNOWN_BUGCHECK_CODE (0xA1000005), the name of the failing function (amdfender.sys), and the reason for the failure; amdfendr.sys called an invalid function - that means that it did something that it's not allowed to do but we don't know exactly what. However, because these are live kernel dumps and not BSOD dumps we know that Windows managed to recover, probably by crashing the process making the graphics calls. That could lead to the black screen that you're seeing.

I suggest you either look for an update to the AMD Crash Detection System or (better still) fully uninstall it. Windows contains it's own graphics hang detection and recovery system so you really don't need the AMD one.
 
S

Swinny

Bronze Level Poster
ubuysa said:
Thanks for those, they do make thinks a lot clearer. The dumps are all identical (as I expected) and all of them have an UNKNOWN_BUGCHECK_CODE (0xA1000005) which means that the kernel detected an error but wasn't able to categorise what type of error it was.

The problem in all the dumps is with amdfendr.sys, the AMD Crash Defender System. This is supposed to recover from hangs in your AMD graphics card and/or driver. In the dumps we see the AMD graphics driver (amdkmdag.sys) called repeatedly, which is normal and expected, but then we see amdfendr.sys called, that indicates that a graphics hang was detected by the AMD Crash Detection System and recovery was attempted. You can see this in an extract of the function call stack (which you read from the bottom up)...
Code: 
ffffb688`e9f2e818  fffff804`c16c34d8 amdfendr+0x534d8
ffffb688`e9f2e820  00000000`0000027f
ffffb688`e9f2e828  00000000`00000000
ffffb688`e9f2e830  00000000`00000000
<empty stack records skipped>
ffffb688`e9f2ec38  fffff804`c16874ba amdfendr+0x174ba
ffffb688`e9f2ec40  00000000`00000000
ffffb688`e9f2ec48  00000000`00000000
ffffb688`e9f2ec50  00000000`00000000
ffffb688`e9f2ec58  00000000`00000000
ffffb688`e9f2ec60  00000000`00000000
ffffb688`e9f2ec68  00000000`00000000
ffffb688`e9f2ec70  ffffb688`e9f2ed40
ffffb688`e9f2ec78  00000000`00000000
ffffb688`e9f2ec80  ffff60f3`cd37c3c0
ffffb688`e9f2ec88  ffff938c`559b08d0
ffffb688`e9f2ec90  00000000`00000000
ffffb688`e9f2ec98  00000000`00000008
ffffb688`e9f2eca0  00000000`00000000
ffffb688`e9f2eca8  00000000`00000006
ffffb688`e9f2ecb0  00000000`00000008
ffffb688`e9f2ecb8  00000000`00000000
ffffb688`e9f2ecc8  fffff804`c1674b33 amdfendr+0x4b33
ffffb688`e9f2ecd0  ffff938c`4b913000
ffffb688`e9f2ecd8  00000000`00000000
ffffb688`e9f2ece0  fffff804`dce293c0 amdkmdag+0x14093c0
ffffb688`e9f2ece8  fffff804`dbefc415 amdkmdag+0x4dc415
You see at the bottom the amdkmdag.sys graphics driver being called a couple of times at different offsets. This is perfectly normal behaviour. Then we see the first amdfendr.sys call and this must be because the AMD Crash Detection System detected (or thought it detected) a graphics hang. Then we see a second call the amdfendr.sys at a different offset, followed by a third call at a different offset again and that's where the function call stack stops, so the error was in that third amdfender.sys function call.

This is confirmed by the dump triage analysis which also points at amdfendr.sys...
Code: 
FAILURE_BUCKET_ID:  LKD_0xA1000005_amdfendr!unknown_function
There you see LKD meaning this is a live kernel dump, the UNKNOWN_BUGCHECK_CODE (0xA1000005), the name of the failing function (amdfender.sys), and the reason for the failure; amdfendr.sys called an invalid function - that means that it did something that it's not allowed to do but we don't know exactly what. However, because these are live kernel dumps and not BSOD dumps we know that Windows managed to recover, probably by crashing the process making the graphics calls. That could lead to the black screen that you're seeing.

I suggest you either look for an update to the AMD Crash Detection System or (better still) fully uninstall it. Windows contains it's own graphics hang detection and recovery system so you really don't need the AMD one.
Click to expand...
Thank you for all of this and your help.

Even though alot of it goes over my head haha.

How would one go around uninstalling this if you think this would be the best process?

Thanks again
 
ubuysa

ubuysa

The BSOD Doctor
Perhaps as a first step simply disabling the crash defender service would be a wiser option?

Here’s how:
  1. Open the AMD Radeon Software by right-clicking on your desktop and selecting “AMD Radeon Software” from the context menu.
  2. In the Radeon Software interface, click on the “Settings” tab located at the bottom right corner of the window.
  3. Under the “Preferences” section, click on “AMD Crash Defender” to access the settings for the service.
  4. Toggle the switch to disable the AMD Crash Defender Service.
 
SpyderTracks

SpyderTracks

We love you Ukraine
ubuysa said:
Perhaps as a first step simply disabling the crash defender service would be a wiser option?

Here’s how:
  1. Open the AMD Radeon Software by right-clicking on your desktop and selecting “AMD Radeon Software” from the context menu.
  2. In the Radeon Software interface, click on the “Settings” tab located at the bottom right corner of the window.
  3. Under the “Preferences” section, click on “AMD Crash Defender” to access the settings for the service.
  4. Toggle the switch to disable the AMD Crash Defender Service.
Click to expand...
IMHO, the crash defender is not the issue, the issue is why it's crashing, as it works absolutely fine throughout 99% of systems. There will be something conflicting with it causing the driver timeouts.

Although if that method does work, it's a quick and painless fix rather than trying to root down the offending incompatibility.

But so long as you're following best practice of a clean install each time you install a driver update rather than installing over the top, that you're sure drivers are recent, then it could point to an issue with that latest driver version, or some kind of software incompatibility somewhere.
 
Last edited:
You must log in or register to reply here.
Top