DaVinci Studio 18 Several PC Crashes

I found another interesting thing: you are using bios F37c and there are reports of crashes with that version:
https://www.reddit.com/r/gigabytegaming ... e_screens/

One user in particular wrote:

No bios above F32 is stable with this board. Tried them all and flashed every which way possible to make sure, and cleared, etc. None above F32 are stable. Problem lies in the RAM Profiles XMP. If you enable XMP in any bios above F32, the voltage doesn't increase and that results in crashes, lockups, freezes on startup, no reboot, etc. F32 bios and less, XMP when enabled adjusts the voltages properly based on XMP values.

So while you obviously can boot and work elsewhere without crashes, there might be a problem with the XMP profiles like he wrote.

I saw that there is a newer BIOS F37d from 2022.07.27
I do not think it would be a good idea to go back to BIOS F32, because with F35 they fixed major vulnerabilities. But you could try to revert back to BIOS F37a, too.

BTW, do you have the Re-size bar option for AMD Smart Access Memory enabled? If yes it might be worth to try disabling it. Also read that this could result in instabilities and isn't relevant for anything other than games.

yes, The day before yesterday I updated the bios really hoping in a solution.

EDiT: I have not any XMP active in the BIOS.

Re-size bar active? Uhm I have to check it since those are feature Indon't use ... now the PC is busy with the 3rd pass. I'm writing from tablet

- DANTE VIA: I don't know why it pop-up that version but I have the last version. I checked to be sure but I really have the last one. 1.3.2.1. BTW it works fantastically with DRS 17 and 18 (I used few times with)

That's interesting as I saw the driver version in your System Info file you generated.

Could you check in the device manager the Dante Via device and which driver version it is using?
Sysinfo says it is DANTEVIADRIVER.SYS Version 1.1.2.1 (11/09/2021)

Maybe you have to update that manually because the installer failed to do so?

- DANTE VIA: I don't know why it pop-up that version but I have the last version. I checked to be sure but I really have the last one. 1.3.2.1. BTW it works fantastically with DRS 17 and 18 (I used few times with)

That's interesting as I saw the driver version in your System Info file you generated.

Could you check in the device manager the Dante Via device and which driver version it is using?
Sysinfo says it is DANTEVIADRIVER.SYS Version 1.1.2.1 (11/09/2021)

Maybe you have to update that manually because the installer failed to do so?

I updated it manually since I have the temporary free version ... it expires in days but I will buy it. I]The version number I gave you is the one comes out clicking on it. Strange also it says 11/09/20021 since I installed it almost 1 month ago.

btw: the memtest still running ... it's now 9h57min ... again in the 3rd pass at the moment: no errors

Could you also download the portable version of hwinfo and generate a report (HTML) and upload that?
https://www.hwinfo.com/download/

I'd like to see more about your memory timings, voltage, and so on.

Could you also download the portable version of hwinfo and generate a report (HTML) and upload that?
https://www.hwinfo.com/download/

I'd like to see more about your memory timings, voltage, and so on.

Sure but when it has finished.

I think it will need again 4hours. That memtest is very long

- DANTE VIA: I don't know why it pop-up that version but I have the last version. I checked to be sure but I really have the last one. 1.3.2.1. BTW it works fantastically with DRS 17 and 18 (I used few times with)

That's interesting as I saw the driver version in your System Info file you generated.

Could you check in the device manager the Dante Via device and which driver version it is using?
Sysinfo says it is DANTEVIADRIVER.SYS Version 1.1.2.1 (11/09/2021)

Maybe you have to update that manually because the installer failed to do so?

Memtest finished now (01:30) successfully passed. Not any issue found.
Re-Size BAR is disabled.
Please here the zip with the HTML you asked for

Ryzen.zip

(23.23 KiB) Downloaded 29 times

In this moment: crashed. GPU Temperature: 62°C, CPU Temperature: 63°C.
Core voltage: an average of 1.43V

This only running DaVinci 18

not any other stress test have given failures. I'm getting crazy

i'm desperate I can't even work... I'm getting crazy

I don't wish to be flippant Bart, but given you believe this is an incompatibility issue specifically between Resolve 18 and your hardware (rather than a hardware issue itself), is there a reason you have not reverted to v17 so that you can work in peace, at least until there's a new version you can try, or you can secure access to some alternative components to test on your system? I appreciate its frustrating, but in your situation, I think I'd rather get my work done than limp along with a faulting system.

I saw that your memory is running at low default 2133, despite being rated for 3200.

You might try the following in your BIOS setup:

Enable / load the XMP profile for your memory.

* Set DRAM voltage manually to 1.35V

* Set RAM multiplier to 32x

* Set the first five sub timings manually to 16-18-18-18-36 (the same as XMP would)

* Set the sixth timing (tRC) to 56 (instead of 54 as XMP would)

Geardown mode (GDM) should be enabled for DDR4 >2666

* Set ProcODT to 40 Ohm

Check CPU settings:
FCLK should be in 1:1 sync with the DRAM clock, in that case 1600 MHz
Vcore SOC: 1.1 V (should be on AUTO)
cLDO VDDP: 0.95 V
cLDO VDDG: 0.95 V
VDDG IOD: 1.05 V

You can use this free tool to check if and which timings are in use after booting into Windows:
https://zentimings.protonrom.com

i'm desperate I can't even work... I'm getting crazy

I don't wish to be flippant Bart, but given you believe this is an incompatibility issue specifically between Resolve 18 and your hardware (rather than a hardware issue itself), is there a reason you have not reverted to v17 so that you can work in peace, at least until there's a new version you can try, or you can secure access to some alternative components to test on your system? I appreciate its frustrating, but in your situation, I think I'd rather get my work done than limp along with a faulting system.

I can't, since the Database is modified by 18.
i wanted also to try to install17 in a separate partition, but the database is an issue.
Moreover I have not clue where to get the 17 back (even)

I saw that your memory is running at low default 2133, despite being rated for 3200.

You might try the following in your BIOS setup:

Enable / load the XMP profile for your memory.

* Set DRAM voltage manually to 1.35V

* Set RAM multiplier to 32x

* Set the first five sub timings manually to 16-18-18-18-36 (the same as XMP would)

* Set the sixth timing (tRC) to 56 (instead of 54 as XMP would)

Geardown mode (GDM) should be enabled for DDR4 >2666

* Set ProcODT to 40 Ohm

Check CPU settings:
FCLK should be in 1:1 sync with the DRAM clock, in that case 1600 MHz
Vcore SOC: 1.1 V (should be on AUTO)
cLDO VDDP: 0.95 V
cLDO VDDG: 0.95 V
VDDG IOD: 1.05 V

You can use this free tool to check if and which timings are in use after booting into Windows:
https://zentimings.protonrom.com

I did all what I could and you suggested.

and before to answer here, I started to edit a video to see if everything was ok with. After a while: sudden switch off. Core Temperature 62°C GPU 71°C

EDIT: just to give another datil: when I use OBS, I reach same temperatures, but I don't get any fault/crash/switch off

I can't, since the Database is modified by 18.
i wanted also to try to install17 in a separate partition, but the database is an issue.
Moreover I have not clue where to get the 17 back (even)

You can consider exporting your working project timelines using XML, then import those to a fresh v17 database / project (or to your previous v17 database backup from prior to your v18 update, if available). You can download Resolve 17.4.6 (or any prior version) from the support page.

By the way, not sure what particular Ryzen 7 processor you're using (nor your specific PSU) but if you happen to be running a Ryzen 5000 series processor then (as suggested earlier in this thread) you might want to look at testing your undervolting options. There's a well presented YouTube review of using Precision Boost Overdrive 2 to that end here:

I can't, since the Database is modified by 18.
i wanted also to try to install17 in a separate partition, but the database is an issue.
Moreover I have not clue where to get the 17 back (even)

You can consider exporting your working project timelines using XML, then import those to a fresh v17 database / project (or to your previous v17 database backup from prior to your v18 update, if available). You can download Resolve 17.4.6 (or any prior version) from the support page.

By the way, not sure what particular Ryzen 7 processor you're using (nor your specific PSU) but if you happen to be running a Ryzen 5000 series processor then (as suggested earlier in this thread) you might want to look at testing your undervolting options. There's a well presented YouTube review of using Precision Boost Overdrive 2 to that end here:

Hello Andy, for the DAVinci versions: I just searched but not found anything rather than the 18. Perhaps I searched in the wrong way.

I will look again later. Perhaps I can install it on another disk avoiding to use PostgreSQL so that it doesn't interfere with. And I can try to work with for the moment in time.

About the Ryzen
I have the AMD Ryzen 7 3700X

As I said, it's not a real crash. not any BSOD ... just the PC switches off ... something that doesn't occur with any other application

For Resolve Studio 17.4.6 (or earlier) go to : https://www.blackmagicdesign.com/support/
then scroll down in the Latest Downloads section until you see the version you're looking for:

Screenshot 2022-08-18 at 14.57.08.jpg (385.59 KiB) Viewed 1906 times

For Resolve Studio 17.4.6 (or earlier) go to : https://www.blackmagicdesign.com/support/
then scroll down in the Latest Downloads section until you see the version you're looking for:

Screenshot 2022-08-18 at 14.57.08.jpg

yes right. I just found it im' downloading it.

How I have to rename the appdata/blackmagicdesign so that it wlll be not polluted
And I have to avoid to use PostreSQL to avoid disruptive interferences.

I renamed
\AppData\Roaming\Blackmagic Design
\AppData\Roaming\Blackmagic Design.18

the 17 is installed and ... it has created a new
\AppData\Roaming\Blackmagic Design

it tells that the database is not compatible.
and when I go to check the DB files, it has the exact structure I had in the 18

into the new created directory: \AppData\Roaming\Blackmagic Design


If I know where di d ti copied it I can do something ora I can't do anything

not easy to fal lback however Projects created with 18 bye-bye but in this case I can easily replicate the last one. It's required to spot where the problem starts

I don't think this is CPU/GPU temperature related. There must be another variable involved we haven't thought of yet. Did you build the PC yourself? Are there any power cable extensions (inside your case I mean) in use besides the ones you got with the power supply? Check inside if power cables into the power supply (if it got modular cabling) and into devices and mobo are seated properly.

Can you re-seat your memory modules and SATA cables?
Then do a new test in Resolve.

When CPU and GPU are under load in Resolve and memory gets also under load - maybe the memory gets not enough cooling. How is the air flow in your PC case? Are there fans on the front blowing inwards and a fan on the back blowing outwards?

Can you remove any device connected to USB (audio stuff and Speed Editor) besides mouse and keyboard? Just to take those out of the equation. Maybe take a photo before unplugging so you remember where each device was plugged in.
Then do a new test in Resolve.

I don't think this is CPU/GPU temperature related. There must be another variable involved we haven't thought of yet. Did you build the PC yourself? Are there any power cable extensions (inside your case I mean) in use besides the ones you got with the power supply? Check inside if power cables into the power supply (if it got modular cabling) and into devices and mobo are seated properly.

Can you re-seat your memory modules and SATA cables?
Then do a new test in Resolve.

When CPU and GPU are under load in Resolve and memory gets also under load - maybe the memory gets not enough cooling. How is the air flow in your PC case? Are there fans on the front blowing inwards and a fan on the back blowing outwards?

Can you remove any device connected to USB (audio stuff and Speed Editor) besides mouse and keyboard? Just to take those out of the equation. Maybe take a photo before unplugging so you remember where each device was plugged in.
Then do a new test in Resolve.

In this moment I'm working with the 17.4.6 ... I rebuilt the episode .... I wil work till the moment crash .. in both cases I will be back to report.

Ok here the report:

Suddenly also with the 17.4.6 shutted off.
then the issue is not DRS18 with itself.

The only one thing now to imagine, is some update with Windows 11 that under DRS causes this issue probably with this HW I have

This is more frustrating. Because I worked perfectly with the 17.4.6 till 18 was released. I finished one logn episode for a client with the 17.4.6 before to update to 18

But then I didn't use for a while (holiday time) meanwhile I had some W11 Updates yes.
I also experimented a bit the 18 trying to see if really Fusion whas more flud

Till that time: no issues.

Then I started to make the episode (a tutorial Short) I sent to Dwaine but this is just for fun and so I could check the DRS18 during a real montage.
then I started to get this issue

if somebody can tast this under my exact conditions (HW and OS) will be great.

From the LOG ERROR
Critical:

Code: Select all

- System

  - Provider

   [ Name]  Microsoft-Windows-Kernel-Power
   [ Guid]  {331c3b3a-2005-44c2-ac5e-77220c37d6b4}
 
   EventID 41
 
   Version 8
 
   Level 1
 
   Task 63
 
   Opcode 0
 
   Keywords 0x8000400000000002
 
  - TimeCreated

   [ SystemTime]  2022-08-18T14:49:01.6995012Z
 
   EventRecordID 114683
 
   Correlation
 
  - Execution

   [ ProcessID]  4
   [ ThreadID]  8
 
   Channel System
 
   Computer ROOT-HP
 
  - Security

   [ UserID]  S-1-5-18
 

- EventData

  BugcheckCode 0
  BugcheckParameter1 0x0
  BugcheckParameter2 0x0
  BugcheckParameter3 0x0
  BugcheckParameter4 0x0
  SleepInProgress 0
  PowerButtonTimestamp 0
  BootAppStatus 0
  Checkpoint 0
  ConnectedStandbyInProgress false
  SystemSleepTransitionsToOn 0
  CsEntryScenarioInstanceId 0
  BugcheckInfoFromEFI false
  CheckpointStatus 0
  CsEntryScenarioInstanceIdV2 0
  LongPowerButtonPressDetected false


ERROR:

Code: Select all

- System

  - Provider

   [ Name]  EventLog
 
  - EventID 6008

   [ Qualifiers]  32768
 
   Version 0
 
   Level 2
 
   Task 0
 
   Opcode 0
 
   Keywords 0x80000000000000
 
  - TimeCreated

   [ SystemTime]  2022-08-18T14:49:10.4712113Z
 
   EventRecordID 114665
 
   Correlation
 
  - Execution

   [ ProcessID]  0
   [ ThreadID]  0
 
   Channel System
 
   Computer ROOT-HP
 
   Security
 

- EventData

   16:35:45
   ‎08/‎18/‎2022
   
   
   4811
   
   
   E607080004001200100023002D00AE02E6070800040012000E0023002D00AE02600900003C000000010000006009000001000000B00400000000000000000000


--------------------------------------------------------------------------------

Binary data:


In Words

0000: 000807E6 00120004 00230010 02AE002D
0010: 000807E6 00120004 0023000E 02AE002D
0020: 00000960 0000003C 00000001 00000960
0030: 00000001 000004B0 00000000 00000000


In Bytes

0000: E6 07 08 00 04 00 12 00   æ.......
0008: 10 00 23 00 2D 00 AE 02   ..#.-.®.
0010: E6 07 08 00 04 00 12 00   æ.......
0018: 0E 00 23 00 2D 00 AE 02   ..#.-.®.
0020: 60 09 00 00 3C 00 00 00   `...<...
0028: 01 00 00 00 60 09 00 00   ....`...
0030: 01 00 00 00 B0 04 00 00   ....°...
0038: 00 00 00 00 00 00 00 00   ........


It seems increasing likely your Power Supply may be developing / have developed a fault, Bart.

Quoting from https://docs.microsoft.com/en-us/window ... 41-restart

Event ID 41 isn't recorded or the bug check code is zero. This behavior might indicate a power supply problem... The following conditions might be the cause:
In the case of a portable computer, the battery was removed or drained.
In the case of a desktop computer, the computer was unplugged or experienced a power outage.
The power supply is underpowered or faulty.

Typically, the symptoms described in this scenario indicate a hardware problem. To help isolate the problem, do the following steps:

Disable overclocking. If the computer has overclocking enabled, disable it. Verify that the issue occurs when the system runs at the correct speed.
Check the memory. Use a memory checker to determine the memory health and configuration. Verify that all memory chips run at the same speed and that every chip is configured correctly in the system.
Check the power supply. Verify that the power supply has enough wattage to appropriately handle the installed devices. If you added memory, installed a newer processor, installed more drives, or added external devices, such devices can require more energy than the current power supply can provide consistently. If the computer logged Event ID 41 because the power to the computer was interrupted, consider obtaining an uninterruptible power supply (UPS) such as a battery backup power supply.
Check for overheating. Examine the internal temperature of the hardware and check for any overheating components.

It seems increasing likely your Power Supply may be developing / have developed a fault, Bart.

Quoting from https://docs.microsoft.com/en-us/window ... 41-restart

Event ID 41 isn't recorded or the bug check code is zero. This behavior might indicate a power supply problem... The following conditions might be the cause:
In the case of a portable computer, the battery was removed or drained.
In the case of a desktop computer, the computer was unplugged or experienced a power outage.
The power supply is underpowered or faulty.

Typically, the symptoms described in this scenario indicate a hardware problem. To help isolate the problem, do the following steps:

Disable overclocking. If the computer has overclocking enabled, disable it. Verify that the issue occurs when the system runs at the correct speed.
Check the memory. Use a memory checker to determine the memory health and configuration. Verify that all memory chips run at the same speed and that every chip is configured correctly in the system.
Check the power supply. Verify that the power supply has enough wattage to appropriately handle the installed devices. If you added memory, installed a newer processor, installed more drives, or added external devices, such devices can require more energy than the current power supply can provide consistently. If the computer logged Event ID 41 because the power to the computer was interrupted, consider obtaining an uninterruptible power supply (UPS) such as a battery backup power supply.
Check for overheating. Examine the internal temperature of the hardware and check for any overheating components.

Thank you Andy, I suppose you were able to identify this form the posted codes that to me are totally unknown.

then the Corsair should be under Warranty (very long one) so I will check it.
Unfortunately i have not any instrument to test it (a dynamic dummy lad etc)
but I will follow that path as well ...

Im' wondering why it doesn't collapse under all the tests I have made under the suggestions of Robert .. especially the 1st test that was putting it under stress

Bart, I think you haven't mention which power supply exactly you are having, have you?

For some power supplies it is important to balance the load of the 12 V rail, this should be mentioned in the power supply's manual. So it might be necessary to connect the motherboard 12 V and the GPU 12 V on different rails.

Also check that your Radeon FX590 is connected to both 8 and 6-pin power cables.

For instance now it's under PowerMAX and developed temperatures 77°C GPU 79°C CPU and doesn't switch off ...

Bart, I think you haven't mention which power supply exactly you are having, have you?

For some power supplies it is important to balance the load of the 12 V rail, this should be mentioned in the power supply's manual. So it might be necessary to connect the motherboard 12 V and the GPU 12 V on different rails.

Also check that your Radeon FX590 is connected to both 8 and 6-pin power cables.

The power supply is Corsair HX750i

yes the GPU had both connectors connected (the 8 and the 6 pins)

It is that version of DaVinci Resolve that due to some incompatibility in some situation trigger the whole system crash. There is no other explanation. It is clear that the system is stable when pushed by other software.
Is there any feedback in event viewer?

It is that version of DaVinci Resolve that due to some incompatibility in some situation trigger the whole system crash. There is no other explanation. It is clear that the system is stable when pushed by other software.
Is there any feedback in event viewer?

The only feedback I got I published here:

viewtopic.php?f=21&t=165696&p=876348#p876327

now it's 20 minutes is running under powerMAX and the PSU is not collapsing at all ...

CPU Temperature: 81°C GPU Temperature 79°C ... But when goes under DRS with much lower temperatures, all of sudden it turns off and./or resets (the PC restarts)

Resolve may be causing power spikes due to ramping demands on your hardware that is causing your PSU to fault... thats why folks were suggesting you might try undervolting your system, to reduce demands including spikes. That said, I'm not a hardware guru, so you'd be better listening to folks who can offer a deeper understanding than me. Still, if you have access to any spare components, especially an alternative PSU that you can test with your PC, thats what I would be trying. Perhaps you have friends, colleagues or family that might be willing to let you borrow their power supply for couple of hours? Failing that you might consider arranging to test with a more capable PSU at a local PC repair shop. Fwiw, mine is an 860W single rail unit.

I'll throw this out. See if you can find a gaming benchmark program. Likely the closest you'll get to the sort of load Resolve might be creating.

Also put up task manager and watch it. See if the system is maxing the GPU or CPU just before it crashes. I know this won't be easy.

Just to be clear: this is a hardware or system-layer problem on that one machine. There is nothing wrong with Resolve and there is no Resolve source code that can be changed to avoid a problem with a single specific machine.

The application layer places a complex, multi-faceted load on the OS and system layers below it. That load could change from version to version but it's the sole responsibility of the lower layers to handle that. Seeking resolution at the application layer for a machine or OS crash is typically not productive. It tends to waste lots of time in "wild goose chases".

As hardware systems become more complex it gets progressively harder for synthetic tests to simulate a specific application workload. That workload is always changing, and not just from source code differences.

E.g, a new app version might use a different compiler optimization that exercises vector instructions, causing the vector units on some CPUs to burn more power. Or the developer could use profile-guided optimization which improves CPU cache hit ratio, but the CPU spends less time waiting on memory so consumes more power.

That is why large PC manufacturers have integration and test departments. Before the complete machine is marketed and sold, some engineer sits in front of a logic analyzer studying an infrequent, transient glitch caused by an incompatibility between components:

https://joema.smugmug.com/Computers/Tek ... 52kXRLG/XL

They send trace files to their counterparts at other companies trying to figure out whose component or whose firmware is responsible for some difficult-to-reproduce problem. When you get a machine like an HP Z workstation or a Mac, that integration troubleshooting has been done. When you build the PC yourself, that has not been done to the same degree. I've built many PCs and the results can be good but sometimes painstaking troubleshooting can be required. Usually a software application vendor is not optimally equipped to take a leading role in debugging problems at the system and hardware layers which are out of their control.

I'll throw this out. See if you can find a gaming benchmark program. Likely the closest you'll get to the sort of load Resolve might be creating.

Also put up task manager and watch it. See if the system is maxing the GPU or CPU just before it crashes. I know this won't be easy.

I got Heaven Benchmark 4.0 that make the driver of the FX590 hung, after a while.
Like DaVinci does sometimes while rendering.
This I know it's a bug of AMD. I sent reports several times.

But the PC still running ...

I suppose that Ryzen 7 have a GPU itself. Maybe you can try this without the RX590.

Just to be clear: this is a hardware or system-layer problem on that one machine. There is nothing wrong with Resolve and there is no Resolve source code that can be changed to avoid a problem with a single specific machine.

The application layer places a complex, multi-faceted load on the OS and system layers below it. That load could change from version to version but it's the sole responsibility of the lower layers to handle that. Seeking resolution at the application layer for a machine or OS crash is typically not productive. It tends to waste lots of time in "wild goose chases".

As hardware systems become more complex it gets progressively harder for synthetic tests to simulate a specific application workload. That workload is always changing, and not just from source code differences.

E.g, a new app version might use a different compiler optimization that exercises vector instructions, causing the vector units on some CPUs to burn more power. Or the developer could use profile-guided optimization which improves CPU cache hit ratio, but the CPU spends less time waiting on memory so consumes more power.

That is why large PC manufacturers have integration and test departments. Before the complete machine is marketed and sold, some engineer sits in front of a logic analyzer studying an infrequent, transient glitch caused by an incompatibility between components:

https://joema.smugmug.com/Computers/Tek ... 52kXRLG/XL

They send trace files to their counterparts at other companies trying to figure out whose component or whose firmware is responsible for some difficult-to-reproduce problem. When you get a machine like an HP Z workstation or a Mac, that integration troubleshooting has been done. When you build the PC yourself, that has not been done to the same degree. I've built many PCs and the results can be good but sometimes painstaking troubleshooting can be required. Usually a software application vendor is not optimally equipped to take a leading role in debugging problems at the system and hardware layers which are out of their control.

ok now is more clear, even if it's difficult to understand what's occurring in a project (1 minute video) where there are just 2 footages, 3 soundtracks, a subtitle and 1 title. No VFX or color correction at all. It doesn't seem respurces demanding.
However tomorrow I will fill the RMA with Corsair ... I have to unmount the PSU to get the pic of the serial (since they expressely ask for it in this way) and then I hope with the new one evrything will be fine. It's now a Marathon

I feel with you. Such a problem is just horrible.
It would be easier if I could be physically at your place to examine the computer.
Cudos to your patience going through all the tests I had to throw at you to single out the issue.

I feel with you. Such a problem is just horrible.
It would be easier if I could be physically at your place to examine the computer.
Cudos to your patience going through all the tests I had to throw at you to single out the issue.

A big thank and cudos to all of you because you helped me to identify the root of this sneaky issue.

I downloaded OCCT and I'm watching fluctuations.

the 12V is stable

GPU Core Voltage (VDCC) is 0.950 on any application. While under DRS it cosa up to 1.250
Core 0VID 1.39-1.40V

As I said with any other application even 3D CAD etc: it stays down.

EDIT: testing now OBS. 0.950V In recording: 1.1V

With DRS: wheil it's opening the Control panel: raises to 1.25 than goes back. As soon as I open a project .. just opening it: raises to stable 1.250

Check CPU settings:
FCLK should be in 1:1 sync with the DRAM clock, in that case 1600 MHz
Vcore SOC: 1.1 V (should be on AUTO)
cLDO VDDP: 0.95 V
cLDO VDDG: 0.95 V
VDDG IOD: 1.05 V

Hi Robert

In your suggestions, you gave me also these 2 parameters, that honestly I don't see on my BIOS.
Can you help me with a screenshot?

Because the VDDG and VDDP the only ones I found are or Auto or Manual and on Manual I can't insert decimals ... just integers so i'm not sure those are the parameters you suggested

Thank you

I'll have a look in my PC's BIOS and make you some screenshots.

I'll have a look in my PC's BIOS and make you some screenshots.

Thank you

I'll have a look in my PC's BIOS and make you some screenshots.

Hi Robert

I found a video where it's shown.

It is for the Extreme, while I have the Elite, but they are pretty similar

One thing important is also to set XMP High Frequency Support from Auto to Level3 (with my type of Memories)

I'm making further tests however, meanwhile Corsair is responding even if slowly, to my RMA Request.

Bart, that's great.
I haven't been able yet to reboot my PC for the BIOS settings as there are some tools running for many hours which I can't stop until finished.

Bart, that's great.
I haven't been able yet to reboot my PC for the BIOS settings as there are some tools running for many hours which I can't stop until finished.

Hi Robert, the PSU is on its trip to Netherlands.
However I spotted something disturbing and not detected on any tests I made. One module on memory bank, has issue with SPD. This can perhaps explain why only after a cold boot. What's not clear: why after a really cold boot (1 night) and not after just 2h. However, in the last days, it started to do it frequentlynso it was easier to test one-by-one the modules to get the strange one. Important: this has nothing to do with the crash I reported in another post. There I posted the report from Windows but it still without answer. Perhaps they think: oh no again that guy woth faulty PSU ...


Sent from my iPad using Tapatalk

Hi,

Did you ever resolve this issue? Was it the PSU? I have a friend who has the same issue (AMD Ryzen 9, RYX 3080). Trying to figure out what's wrong. I told him it might be his PSU, temps etc etc. Then I googled it and found this thread. I read the whole thread, and I'm at the stage now where I think it might be a PSU issue.

Please do let me know if you figured this out, I'd appreciate it.

Thank you so much.

Hi,

Did you ever resolve this issue? Was it the PSU? I have a friend who has the same issue (AMD Ryzen 9, RYX 3080). Trying to figure out what's wrong. I told him it might be his PSU, temps etc etc. Then I googled it and found this thread. I read the whole thread, and I'm at the stage now where I think it might be a PSU issue.

Please do let me know if you figured this out, I'd appreciate it.

Thank you so much.

The PSU was delivered and it wasn't the 'cause.

What caused all the trouble, was a defective memory module.
Not the memory itself. All tests I have made didn't get any error.
It turned out that the defect was due a defective E²PROM on the DDR4 Module.
Each DDR4 has an E²PROM that contains the data of the Memory, used to communicate the bank's features to the UEFI.
Whenever this fails during anytime, or the PC doesn't boot correctly and resets the UFEI/BIOS or you can star tto get very strange behaviour under specific conditions.

Was it easy to find? Not at all. It asked a big troubleshooting work.
Did Corsair replaced the bank? Not at all, because that memory type I need, is not any longer available.
They wanted to send me a bank (2 modules) with slower timings than the other bank (2 modules) I have (in total I have 2 banks (4 modules) on my machine).

From that moment on, I'm working with just 3 modules (1.5 Bank).
So, instead of 64GB i have now 48GB.

Not any issues from the moment I took out the defective module

Hi,

Did you ever resolve this issue? Was it the PSU? I have a friend who has the same issue (AMD Ryzen 9, RYX 3080). Trying to figure out what's wrong. I told him it might be his PSU, temps etc etc. Then I googled it and found this thread. I read the whole thread, and I'm at the stage now where I think it might be a PSU issue.

Please do let me know if you figured this out, I'd appreciate it.

Thank you so much.

The PSU was delivered and it wasn't the 'cause.

What caused all the trouble, was a defective memory module.
Not the memory itself. All tests I have made didn't get any error.
It turned out that the defect was due a defective E²PROM on the DDR4 Module.
Each DDR4 has an E²PROM that contains the data of the Memory, used to communicate the bank's features to the UEFI.
Whenever this fails during anytime, or the PC doesn't boot correctly and resets the UFEI/BIOS or you can star tto get very strange behaviour under specific conditions.

Was it easy to find? Not at all. It asked a big troubleshooting work.
Did Corsair replaced the bank? Not at all, because that memory type I need, is not any longer available.
They wanted to send me a bank (2 modules) with slower timings than the other bank (2 modules) I have (in total I have 2 banks (4 modules) on my machine).

From that moment on, I'm working with just 3 modules (1.5 Bank).
So, instead of 64GB i have now 48GB.

Not any issues from the moment I took out the defective module

Thank you so much for the response, I appreciate it. I'll pass this on to my friend to see if this fixes the issue.

Thanks again!!

Ouch! One of the worst nightmares of a hardware failure.