8 Pack Team Group 16gb 3600 kit Stability issue on Gigabyte x570 Gaming X(Solved with new PSU)

WgizmO · 26 May 2020 at 15:26

Hi,

I did decide to open new thread ans its new memory and would like to carry on my memory issue from previous post here if its ok with admins.

Memory i Use are:

8 Pack Team Group 16gb kit (2x8gb) DDR4 3600 CL 14-15-15-35 1.45v

All 4 are in slots A2, B2 and A1, B1. XMP on,

So far test:
In cinebenchr20 passed no issues
In intelburn 4 runs on high no issues
In howsomething(forgot the name) rendering in 4k no issues
In Frumark after 10 mins no issues gpu temp 40'C which is super low imo.
In aida had issue and cant find what did fail. All I can see is what on pics in the link:

Test 1 aida:

No clue what its mean.

All I did found on internet regarding my aida fail is:

"The hardware failure message specifically means the stress testing module detected a CPU processing error. It may be caused by a CPU execution glitch, a cache issue or if something happened while the data "travelled" from the memory into the CPU. It's not easy to diagnose such issues, and I'm afraid it's not possible to be more specific.

However, please note that the hardware failure message has nothing to do with the measured thermal state, voltage levels or fan speed status of your system. Those measured values are only shown for you as a reference, and when they go out of a specific range, the test will not stop. The test will only stop if the overall system state becomes so critical that a detectable hardware failure occurs. But even that it quite rare, since most unstable systems would throw a BSoD, freeze, or make an instant system restart when a hardware failure occurs. You can hardly ever see a hardware failure message, since the odds of a single CPU processing error is much lower than such a system stability issue that completely crashes and halts the execution pipeline of the CPU (when the CPU simply resets itself along with a sudden system restart).

As for Process Lasso, I'm not sure what exactly does it do that it can manage to interfere with other processes and cause a CPU processing glitch in them, but it sounds quite dangerous, so it would be best if you stayed away from having it running while using the AIDA64 System Stability Test"

Set Fclk in bios to 1800 and running test:

Running memtest86 atm

Jamin280672 · 26 May 2020 at 15:51

WgizmO said:
Hi,

I did decide to open new thread ans its new memory and would like to carry on my memory issue from previous post here if its ok with admins.

Memory i Use are:

8 Pack Team Group 16gb kit (2x8gb) DDR4 3600 CL 14-15-15-35 1.45v

All 4 are in slots A2, B2 and A1, B1. XMP on,

So far test:
In cinebenchr20 passed no issues
In intelburn 4 runs on high no issues
In howsomething(forgot the name) rendering in 4k no issues
In Frumark after 10 mins no issues gpu temp 40'C which is super low imo.
In aida had issue and cant find what did fail. All I can see is what on pics in the link:

Test 1 aida:

No clue what its mean.

All I did found on internet regarding my aida fail is:

"The hardware failure message specifically means the stress testing module detected a CPU processing error. It may be caused by a CPU execution glitch, a cache issue or if something happened while the data "travelled" from the memory into the CPU. It's not easy to diagnose such issues, and I'm afraid it's not possible to be more specific.

However, please note that the hardware failure message has nothing to do with the measured thermal state, voltage levels or fan speed status of your system. Those measured values are only shown for you as a reference, and when they go out of a specific range, the test will not stop. The test will only stop if the overall system state becomes so critical that a detectable hardware failure occurs. But even that it quite rare, since most unstable systems would throw a BSoD, freeze, or make an instant system restart when a hardware failure occurs. You can hardly ever see a hardware failure message, since the odds of a single CPU processing error is much lower than such a system stability issue that completely crashes and halts the execution pipeline of the CPU (when the CPU simply resets itself along with a sudden system restart).

As for Process Lasso, I'm not sure what exactly does it do that it can manage to interfere with other processes and cause a CPU processing glitch in them, but it sounds quite dangerous, so it would be best if you stayed away from having it running while using the AIDA64 System Stability Test"

Set Fclk in bios to 1800 and running test:

Running memtest86 atm

Gaming ???? It would be the first thing I tried.

opethdisciple · 26 May 2020 at 16:13

Are you running Process Lasso? It's not clear from your description. I would remove it. Unnesecary.

To test stability, download the dram calc, go to the membench section and run a stability test.

These settings:

memtest
400%
single

And let it run. Will take around 2hrs maybe faster. It should not stop with any errors.

WgizmO · 26 May 2020 at 16:14

OK, all tests passed with cflk passed.
Now I didnt left aida stability test for longer than 5 mins as wanted to test games.

1st game I did run was Gears Tactic. Around 20 sec in game menu PC did shutdown same way as it was on other memory set. Difference so far I can se is my bios read xmp no problems not like other kit.

postmanfw · 26 May 2020 at 16:19

WgizmO said:
PC did shutdown same way as it was on other memory set.

Pc shut down completely rather than just crash and restarted?

Thats normally an indication of a power problem.

WgizmO · 26 May 2020 at 16:26

postmanfw said:
Pc shut down completely rather than just crash and restarted?

Thats normally an indication of a power problem.

Yes its shutting down/turning down my pc completely. When I start it again my bios is default.

From what I read on internet wrong memory settings might make mobo to shutdown pc to protect components from dmg but not sure if its thrue.

New PSU arrive tomorrow then will see if its happen with new PSU so far trying to check if ram stable.

opethdisciple said:
Are you running Process Lasso? It's not clear from your description. I would remove it. Unnesecary.

To test stability, download the dram calc, go to the membench section and run a stability test.

These settings:

memtest
400%
single

And let it run. Will take around 2hrs maybe faster. It should not stop with any errors.

membench test finished with scores:

153.67 and best 108.5

Now running memtest with task scope 400 and stop at (task mode) single

memtest86 you mean?

I did run it for aver hour and no errors.

Joxeon · 26 May 2020 at 16:27

WgizmO said:
Hi,

I did decide to open new thread ans its new memory and would like to carry on my memory issue from previous post here if its ok with admins.

Memory i Use are:

8 Pack Team Group 16gb kit (2x8gb) DDR4 3600 CL 14-15-15-35 1.45v

All 4 are in slots A2, B2 and A1, B1. XMP on,

So far test:
In cinebenchr20 passed no issues
In intelburn 4 runs on high no issues
In howsomething(forgot the name) rendering in 4k no issues
In Frumark after 10 mins no issues gpu temp 40'C which is super low imo.
In aida had issue and cant find what did fail. All I can see is what on pics in the link:

Test 1 aida:

No clue what its mean.

All I did found on internet regarding my aida fail is:

"The hardware failure message specifically means the stress testing module detected a CPU processing error. It may be caused by a CPU execution glitch, a cache issue or if something happened while the data "travelled" from the memory into the CPU. It's not easy to diagnose such issues, and I'm afraid it's not possible to be more specific.

However, please note that the hardware failure message has nothing to do with the measured thermal state, voltage levels or fan speed status of your system. Those measured values are only shown for you as a reference, and when they go out of a specific range, the test will not stop. The test will only stop if the overall system state becomes so critical that a detectable hardware failure occurs. But even that it quite rare, since most unstable systems would throw a BSoD, freeze, or make an instant system restart when a hardware failure occurs. You can hardly ever see a hardware failure message, since the odds of a single CPU processing error is much lower than such a system stability issue that completely crashes and halts the execution pipeline of the CPU (when the CPU simply resets itself along with a sudden system restart).

As for Process Lasso, I'm not sure what exactly does it do that it can manage to interfere with other processes and cause a CPU processing glitch in them, but it sounds quite dangerous, so it would be best if you stayed away from having it running while using the AIDA64 System Stability Test"

Set Fclk in bios to 1800 and running test:

Running memtest86 atm

I had Aida fail like this when FCLK wasn't stable. It would fail almost immediately.
Bear in mind I was trying to run 1900 but had to drop it back to 1867 then it ran fine, also a more recent Bios update fixed this and I can now run 1900 without issue.
maybe its just that your chip isn't happy at 1800 so you could try dropping it down a notch and retesting.

opethdisciple · 26 May 2020 at 16:41

WgizmO said:
Yes its shutting down/turning down my pc completely. When I start it again my bios is default.

From what I read on internet wrong memory settings might make mobo to shutdown pc to protect components from dmg but not sure if its thrue.

New PSU arrive tomorrow then will see if its happen with new PSU so far trying to check if ram stable.

membench test finished with scores:

153.67 and best 108.5

Now running memtest with task scope 400 and stop at (task mode) single

memtest86 you mean?

I did run it for aver hour and no errors.

Nope. Read what I wrote closely. Dram calc, membench section there is a memtest from the drop down menu.

Also what CPU are you running?

WgizmO · 26 May 2020 at 16:47

opethdisciple said:
Nope. Read what I wrote closely. Dram calc, membench section there is a memtest from the drop down menu.

Also what CPU are you running?

3900X

opethdisciple · 26 May 2020 at 17:44

WgizmO said:
3900X

That looks better.

Although possible, I doubt your FCLK of 1800 is too much for the 3900x.

From your timings it looks like you will have to spend some time and overclock that ram as they are running in a default configuration and you can get much better performance from a manual overclock.

First of all find out if all is table at stock settings. So once the memtest finishes play some games for a bit. 1hr should suffice. Something CPU/memory demanding.

Then you need to find out what your max FCLK is. Because it might be capable of more than 1800MHz. You should find the FCLK is somewhere between 1800MHz and 1900MHz. Once you know that take your FCLK and * 2 to get you dram speed.

Then use the dram calc specifically version 1.7.0 (not 1.7.3 as it seems to give slower timings) and use the primary and secondary timings plus dram voltage (you may need more than the dram voltage it suggests) to overclock the ram.

Then use the memtest default preset to bench the ram, with a lower score being better than a higher one.

Good luck.

Also this is what we've have been doing in this thread.

WgizmO · 26 May 2020 at 18:57

opethdisciple said:
That looks better.

Although possible, I doubt your FCLK of 1800 is too much for the 3900x.

From your timings it looks like you will have to spend some time and overclock that ram as they are running in a default configuration and you can get much better performance from a manual overclock.

First of all find out if all is table at stock settings. So once the memtest finishes play some games for a bit. 1hr should suffice. Something CPU/memory demanding.

Then you need to find out what your max FCLK is. Because it might be capable of more than 1800MHz. You should find the FCLK is somewhere between 1800MHz and 1900MHz. Once you know that take your FCLK and * 2 to get you dram speed.

Then use the dram calc specifically version 1.7.0 (not 1.7.3 as it seems to give slower timings) and use the primary and secondary timings plus dram voltage (you may need more than the dram voltage it suggests) to overclock the ram.

Then use the memtest default preset to bench the ram, with a lower score being better than a higher one.

Good luck.

Also this is what we've have been doing in this thread.

Atm I would like to have it stable on settings out of the box 3600 cl14. Then when will run stable for week or so can start playing around to get 1:1 my ram if possible.

I did run your test for 140+ minutes and no error. Finished on 200 tasks form 400 as it would take forever not around 2h

Now will try some games. Gears wars crashing straight away then lets say this is a game issue. Will check in WoW. If wow will crash then issue still exist but ram looks stable from all benchmarks I did so far.

opethdisciple · 26 May 2020 at 19:19

If your applications crash straight away when your memory is stable then it could indeed be the FCLK.

From the dram calc apply the VDDG and VDDP voltages it suggests. (This is to stabilise the FCLK)

I assume you know to import your timings first from taiphoon burner to do this.

WgizmO · 26 May 2020 at 19:39

opethdisciple said:
If your applications crash straight away when your memory is stable then it could indeed be the FCLK.

From the dram calc apply the VDDG and VDDP voltages it suggests. (This is to stabilise the FCLK)

I assume you know to import your timings first from taiphoon burner to do this.

Please tell me if I did it correct:

I'm noob at this. I do have 4 ram sticks btw 4x8 GB

Update:

Gears Tactic switch off Pc few just in the menu.
World of Warcraft switch off Pc after 5 minutes of game.

My current Bios settings are:

With this settings all benchmarks passes with no errors. memtest after 2H+ no errors. Games where switch off pc.

opethdisciple · 26 May 2020 at 20:01

No thats not right.

You have to go to taiphoon burner, choose one of the sticks from the menu, go to report, scroll down to where it says 'display in nanoseconds' and then export the complete html profile and finally import to dram calc.

Joxeon · 26 May 2020 at 20:02

WgizmO said:
Please tell me if I did it correct:

You need to convert to nano seconds then save to HTML

Check this post as it shows you all the steps.

https://www.reddit.com/r/Amd/comments/cw3cn0/psa_use_thaiphoon_to_exportimport_your_xmp/

WgizmO · 26 May 2020 at 20:17

Joxeon said:
You need to convert to nano seconds then save to HTML

Check this post as it shows you all the steps.

https://www.reddit.com/r/Amd/comments/cw3cn0/psa_use_thaiphoon_to_exportimport_your_xmp/

opethdisciple said:
No thats not right.

You have to go to taiphoon burner, choose one of the sticks from the menu, go to report, scroll down to where it says 'display in nanoseconds' and then export the complete html profile and finally import to dram calc.

When I open Thaiphoon Burner only option I have is read to chose stick. Then I can do Report. When clicking Report I have few options Report HTML and Complete Report HTML. Ofc before that i did scroll down to click "display in nanoseconds". Then I chose Complete HTML Report. Opening DRAM Calculator, press Import XMP and pressed Calculate Safe:

I did that properly?

WgizmO · 26 May 2020 at 20:22

Now I forgot to chose 4 modules. Now below are 4 Modules:

FAST:

SAFE:

Joxeon · 26 May 2020 at 20:36

Start off with safe. try up to 1.45 Dram volts with this.

WgizmO · 26 May 2020 at 21:49

Joxeon said:
Start off with safe. try up to 1.45 Dram volts with this.

I can't find all settings in bios or can't change them. On below pictures is all what I could find:

Below settings are not working from safe settings. When Pc restart itself it take longer than usual and boot to windows but ram is set as default.

Can't change above values

Can't change above values

Can't find:

Power down mode
BGS
BSS alt
VDDG CCD Voltage
VDDG IOD Voltage
CLDO VDDP Voltage

Don't know what is:

TRFC 2 and 4
TRCPAGE

Sorry guys but I'm new in to this and trying to understand what I'm doing and learn as much as I can.

Joxeon · 26 May 2020 at 22:56

WgizmO said:
Can't find:

Power down mode
BGS
BSS alt
VDDG CCD Voltage
VDDG IOD Voltage
CLDO VDDP Voltage

Have a look in system and go through every setting as theres usually more than one place to adjust all the memory stuff.

Can you not just key in the number here?