Discussion Zen 5 Builders thread

[Markfw]

I think it turns off SMT.

I think it turns off the second CCD for 12 and 16 core CPUs

 

[MarkPost]

Both things: it turns off SMT for all CPUs + the second CCD for 12/16 CPUs

 

[Sonikku]

@Sonikku
Use BlueScreenView and open the dump C:\WINDOWS\Minidump\010425-15343-01.dmp if it still exists.
It should have a stack trace pointing to a module/driver which "caused" the crash.

Thank you for the link here's what it found.

 

[CP5670]

I'll leave it off, don't see any reason to disable SMT.

I'm keeping MCR on this time, since it's on auto (I'm guessing it means on) by default.

 

[igor_kavinski]

here's what it found.

Doesn't show what caused that Windows file to crash.

Event Log Viewer for Windows 11 / 10 / 8 / 7 / Vista

Simple tool for Windows 11/10/8/7/Vista that displays in a table the details of all events from the event logs of Windows, including the event description

www.nirsoft.net

Use that to check the events leading up to the crash. There may be a clue there.

 

[StefanR5R]

For reasons I haven't yet been able to determine, the machine restarts soon after I turn off my display.

Do you also get a crash if you unplug the display cable instead of switching the display off while connected?

Is the display a television set, or a computer monitor? If the former, it might send HDMI-CEC [Wikipedia] commands to the computer when turned off. And a driver misbehaving on reception of these. If so, a driver fix would be desirable, yet if not available, disabling CEC in the TV set could be a workaround.

(Edit: I actually don't know whether or not current graphics cards and graphics drivers feature HDMI-CEC reception, so I may be off on this.)

 

[mmaenpaa]

Thank you for the link here's what it found.

This would indicate video / recording related driver.

Do you have any usb web cams attached?

Or any recording software active, for example have you enabled Radeon ReLive while installing Radeon disply drivers?

 

[Det0x]

K i'm back from my little time away from this forum (have been crazy busy testing new stuff lately)

K its time to show what i have been cooking up lately and the reason why ive been so quiet

I had the lucky opportunity to test the ROG CROSSHAIR X870E APEX a few days before announcement, so i thought i could share my own preliminary results here, but do note this is very far from finalized results




Hardware used for my short tests:

Delidded 9950X ES running PBO
ROG CROSSHAIR X870E APEX
2x16GB G.Skill Trident Z5 Royal Neo A-DIE @ daily 1.76 VDD / 1.55 VDDQ voltages
Very large custom cooling with mora

First off we have some short karhu run, only 1 hour durations do to time constraints.
Core Tuning @ level 2 = over 10mb/s higher throughput then anything i managed on the GENE, ended at a wooping 372.3mb/s


With everything open bandwidth drops alittle, but still much higher than anything else


Also did alittle test in GB3 on the same boot which netted me a wooping 15.2k / 14.75k memory score which is very good for a 9950X i would say
To top it all off, a baby 49k in Cinebench R23 MT while still running PBO



I also did a short 1 hour karhu run @ Core Tuning = legacy -> 53.9ns in AIDA64
Do note its pretty much only in AIDA the latency drops close to 10ns, while Clam latency stays the ~same


All in all i can say from my short test that this board have sick memory performance/effiency, and VDD ceiling seems much better than other boards running agesa 1.2.0.2B
(started a run while i went to work earlier today)

Would recommed if your looking for a memory tweaking board

 

[Sonikku]

This would indicate video / recording related driver.

Do you have any usb web cams attached?

Or any recording software active, for example have you enabled Radeon ReLive while installing Radeon disply drivers?

I have not

Do you also get a crash if you unplug the display cable instead of switching the display off while connected?

Is the display a television set, or a computer monitor? If the former, it might send HDMI-CEC [Wikipedia] commands to the computer when turned off. And a driver misbehaving on reception of these. If so, a driver fix would be desirable, yet if not available, disabling CEC in the TV set could be a workaround.

(Edit: I actually don't know whether or not current graphics cards and graphics drivers feature HDMI-CEC reception, so I may be off on this.)

I think that is the problem! Thank you! I game on an 82" 4k tv and when I disabled HDMI-CEC on the tv itself it stopped causing bluescreens upon turning it off!
...On the other hand, I now need to control my soundbar and tv on seperate remotes. So a driver fix would be preferable for sure. But I'm already on AMD's latest. Any suggestions? I'm guessing I'll need to turn to a 3rd party patch of some kind.

 

[StefanR5R]

Would a DisplayPort-to-HDMI adapter prevent HDMI-CEC from getting through to the PC?
(If you purchase one of these adapters, take care that it is one which is rated for your desired video resolution and refresh rate.)

Edit:
Or search for "HDMI CEC less adapter". These are little plugs or dongles which don't route pin 13. Or take an HDMI cable and isolate pin 13 yourself.

 

[Sonikku]

Tried to install Ryzen Master but I get "ryzen master unable to initialize. Kindly reinstall the program. Hence aborting!" but I get the same message every time I do. Boy you figure out one problem and another emerges. What a nightmare build. XD

Would a DisplayPort-to-HDMI adapter prevent HDMI-CEC from getting through to the PC?
(If you purchase one of these adapters, take care that it is one which is rated for your desired video resolution and refresh rate.)

Edit:
Or search for "HDMI CEC less adapter". These are little plugs or dongles which don't route pin 13. Or take an HDMI cable and isolate pin 13 yourself.

I'll bear it mind. Will consider an adapter as a last resort. I still can't believe something so stupid hasn't been patched/fixed on the drivers side by now.

 

[StefanR5R]

Does the CEC interface show up anywhere in Windows' device manager? If so, try to disable or uninstall it there. The BIOS may also be a place to look, but I am guessing this would only offer HDMI-CEC options for the iGPU, if at all, not for the dGPU.

In theory, instead of a software problem it could also be a hardware problem: Pin 13 of the GPU's HDMI port wired up in a wrong way.

 

[Seba]

Tried to install Ryzen Master but I get "ryzen master unable to initialize. Kindly reinstall the program. Hence aborting!" but I get the same message every time I do. Boy you figure out one problem and another emerges. What a nightmare build. XD

Every time I update Adrenalin graphics driver, Ryzen Master no longer works. It used to be enough to uninstall Ryzen Master, restart, re-install.

But last time this was not enough.
After uninstalling Ryzen Master and rebooting, I also had to delete all registry keys starting with "AMDRyzenMasterDriverVxxx" (there were multiple instances, ending in various numbers) located in "Computer\HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Services".
Then reboot.
Re-install Ryzen Master.
Reboot.
For me it worked.

 

[Hulk]

Absolutely hammering the 9950X right now with a 90 minute Vegas Pro encode using Voukoder to x265, slow setting.
I have the power limit set to 175 Watts because this application hits the cores harder than anything I do in terms of temps.
The die temps seem a little high but the cores seem to only average about 64C.
Anything to worry about here?
These temperatures seems to fluctuate a lot more than my Raptors did. Also the tile temp is so much higher than the core temp?
Am I good at 175W you think or should I back it down a bit?

 

[Markfw]

Absolutely hammering the 9950X right now with a 90 minute Vegas Pro encode using Voukoder to x265, slow setting.
I have the power limit set to 175 Watts because this application hits the cores harder than anything I do in terms of temps.
The die temps seem a little high but the cores seem to only average about 64C.
Anything to worry about here?
These temperatures seems to fluctuate a lot more than my Raptors did. Also the tile temp is so much higher than the core temp?
Am I good at 175W you think or should I back it down a bit?

I think 200 is the default. When I do all cores on heavy avx-512 I fot 80c (stock) so 64c is cool.

No worries in your case.

 

[gdansk]

Also the tile temp is so much higher than the core temp?
Am I good at 175W you think or should I back it down a bit?

Are you trying to reduce the fan noise or the temperature itself?

 

[Hulk]

Are you trying to reduce the fan noise or the temperature itself?

temp

 

[igor_kavinski]

temp

Try:

https://www.reddit.com/r/hardware/comments/zgp2hl/a_huge_comeback_corsairs_xtm70_thermal_paste

 

[Sonikku]

Does the CEC interface show up anywhere in Windows' device manager? If so, try to disable or uninstall it there. The BIOS may also be a place to look, but I am guessing this would only offer HDMI-CEC options for the iGPU, if at all, not for the dGPU.

In theory, instead of a software problem it could also be a hardware problem: Pin 13 of the GPU's HDMI port wired up in a wrong way.

I don't see anything. But maybe it wasn't the issue after all. I had another blue screen...

So many errors... hm. This machine is cursed.

 

[igor_kavinski]

Use the PC for a day or two with just one stick and then again with the other as the sole stick. If the errors happen only with one stick and not the other, that's your issue.

 

[StefanR5R]

Does the CEC interface show up anywhere in Windows' device manager?

I don't see anything.

You looked through all device categories, not just video related ones, right?

But maybe it wasn't the issue after all. I had another blue screen...

In the first screenshot of your three attachments, Ethernet is referenced in the event line which you selected. How about the other lines (records 6062…6095) of the same error event; are they all about networking or also about other subsystems?

And right before that, there are log lines from Task Scheduler. Maybe some background routines (which might also be started after you switched the screen off) are triggering the crashes?

PS, I don't know Windows very well. Besides various Linux PCs I only have an old rarely used Win 7 laptop at home, and I don't administer the Win 10 Dell PC at my workplace myself.

 

[mmaenpaa]

Use the PC for a day or two with just one stick and then again with the other as the sole stick. If the errors happen only with one stick and not the other, that's your issue.

Also one test would be setting memory at default jedec (probably @4800) for a few days. I have had PCs which pass memory tests at expo settings (at least for a few hours) but in "production" had bluescreens once and in a while. After setiing to jedec speed those went away (those were office PC systems so it was not that important to use fastest memory settings)

Currently for office / workstation builds I use fastest jedec (5600) memory. Of course latency settings on those are slow but reliability is number one in business enviroment.

 

[DaaQ]

So started assembling board ect.
Need encouragement.

Remove and repad VRM and backplate VRM pads? I went with Barrow air cooling for RAM.






Also if you remount and or repasted a lot these are really awesome to use.

The isopropyl is 99%. It's for fiber optic cleaning. Just be careful not to scrub too hard.

 

[Hulk]

K its time to show what i have been cooking up lately and the reason why ive been so quiet

I had the lucky opportunity to test the ROG CROSSHAIR X870E APEX a few days before announcement, so i thought i could share my own preliminary results here, but do note this is very far from finalized results
View attachment 114339

View attachment 114340

Hardware used for my short tests:

Delidded 9950X ES running PBO
ROG CROSSHAIR X870E APEX
2x16GB G.Skill Trident Z5 Royal Neo A-DIE @ daily 1.76 VDD / 1.55 VDDQ voltages
Very large custom cooling with mora

First off we have some short karhu run, only 1 hour durations do to time constraints.
Core Tuning @ level 2 = over 10mb/s higher throughput then anything i managed on the GENE, ended at a wooping 372.3mb/s
View attachment 114330

With everything open bandwidth drops alittle, but still much higher than anything else
View attachment 114331

Also did alittle test in GB3 on the same boot which netted me a wooping 15.2k / 14.75k memory score which is very good for a 9950X i would say
To top it all off, a baby 49k in Cinebench R23 MT while still running PBO
View attachment 114332


I also did a short 1 hour karhu run @ Core Tuning = legacy -> 53.9ns in AIDA64
Do note its pretty much only in AIDA the latency drops close to 10ns, while Clam latency stays the ~same
View attachment 114333

All in all i can say from my short test that this board have sick memory performance/effiency, and VDD ceiling seems much better than other boards running agesa 1.2.0.2B
(started a run while i went to work earlier today)
View attachment 114334
Would recommed if your looking for a memory tweaking board

Great information, thanks for sharing.
For that CB 49,000 run, what is your voltage under load? Package power? Frequency? Temps? That's a big CB number!

 

[Det0x]

Great information, thanks for sharing.
For that CB 49,000 run, what is your voltage under load? Package power? Frequency? Temps? That's a big CB number!

That was just a PBO run, around 320w usage i would guess
(wasn't really trying)

Here are full maxed run with more info from benchmate