Ryzen Owners Zen Garden

[AsRock]

Can you elaborate on that?

Some games crash with it on, might be within a minute or a few hours. Red Dead being the biggest offender.

 

[Klemc]

Some games crash with it on, might be within a minute or a few hours. Red Dead being the biggest offender.

Re-BAR on my Z390 BIOS has been pushed and i finally reinstalled older BIOS, no need that "function", 3 games can handle it, if hardware follows...

 

[tpa-pr]

Gigabytes fix doesnt work, as of 1 hour ago with their F10D BIOS
Note that EXPO isn't even enabled here - he still got a 1.7v alert and shutdown via HWinfo, and a screenshot of 1.55v the next event
View attachment 295366


thats how mine behaved, none of my other Zen3 CPU's have worked the same way but the x3D seems to just go -30 and be happy with it.

Load optimised defaults in the BIOS after the CPU change, it helps a lot. That should prevent needing an OS reinstall.

I had this happen to me with my 7700X, actually. No EXPO, no PBO, voltages and temp rocketed for a fraction of a second. Back down to version F9 now and everything is back to normal.

Hoping no lasting damage has been done since it was for a fraction of a second but boy am I a bit upset with Gigabyte. They're trying to make my wife explode.

 

[AusWolf]

Some games crash with it on, might be within a minute or a few hours. Red Dead being the biggest offender.

I've never had that happen. Since you mentioned RDR, I'll have a look once I finish Kingdom Come. Is there any game where it happens faster, and ReBAR is definitely the cause?

Just got my 7800X3D and Gigabyte B650 Elite AX set up this past weekend.

It hits 5000mhz on all cores at 1.2v (latest BIOS applied). Had to manually time the RAM to get it to 6000 mhz- I didn't get the EXPO supported RAM.

Thus far I'm happy. Hearts of Iron 4 and Warhammer 3 are blazing fast.

5000 MHz all core is really nice! That's what my 7700X can do at stock.

 

[Mussels]

idk if it's that simple. What would ASUS gain by keep altering the values higher? Doubling down on destroying more of their own mobos, with or no warranty and get more reddit posts with pics of dead mobos? There must be something more to it.

Probably just pure dumb people doing dumb things, they couldn't get it right.

RAM compatibility, simple as that

SoC voltage is directly linked to how fast you can run RAM on AM4 and AM5

Re-BAR on my Z390 BIOS has been pushed and i finally reinstalled older BIOS, no need that "function", 3 games can handle it, if hardware follows...

If you mean rebar, almost all games benefit from it, and some GPU's suffer massively without it
If you get crashes with it enabled, it's not rebar causing the crashes - it's performing better so your systems working harder and exposing an existing problem

And... zen garden thread. Please keep the intel stuff out.

 

[freeagent]

So, if Zen 4 soc is 1.3v on a smaller node than Zen 3.. is it ok to give Zen 3 1.3v SOC? I know the consensus is 1.2-1.25ish.. but is that just old fears from Zen 2?

I don’t think I ever tried 1.25v.. I probably did

 

[Klemc]

Upgrading from 4650G Pro APU to 5700G will not need any driver change, right (on DeskMini X300) ?

 

[Mussels]

So, if Zen 4 soc is 1.3v on a smaller node than Zen 3.. is it ok to give Zen 3 1.3v SOC? I know the consensus is 1.2-1.25ish.. but is that just old fears from Zen 2?

I don’t think I ever tried 1.25v.. I probably did

1.30v ran hotter with no gains for most people, as the IO die gets unstable at higher temps, so the voltages just made it worse. Past 1.250v i'd say you need to look at other voltages instead.
1.20v ran fine for me, but i'm down to 1.15v now (1.13vish after droop) with slightly lower temps, so it performs better in R23.

Upgrading from 4650G Pro APU to 5700G will not need any driver change, right (on DeskMini X300) ?

You'd want to update the chipset driver and GPU driver, why wouldn't you after any hardware change?

 

[tabascosauz]

Upgrading from 4650G Pro APU to 5700G will not need any driver change, right (on DeskMini X300) ?

Renoir and Cezanne have identical iGPU. But Renoir > Cezanne should do a BIOS update to at least AGESA 1203 if you are not there already, and make sure you are on latest chipset drivers.

 

[Klemc]

You'd want to update the chipset driver and GPU driver, why wouldn't you after any hardware change?

Asrock's DeskMini X300 page for Win10 offers old 3.08.06.148, i have 4.11.15.342 installed (listed in myprogs)...

... will install 5.02.19.2221 after CPU upgrade, and iGPU Adrenalin too, from 23.4.1 to 23.4.3.

Renoir and Cezanne have identical iGPU. But Renoir > Cezanne should do a BIOS update to at least AGESA 1203 if you are not there already, and make sure you are on latest chipset drivers.

I had 1.70 until recently installed 1.80E and after 1.80F (it's always BETA since 1.70+).
QUOTE:

Update AMD AM4 AGESA ComboBIOSv2 PI 1.2.0.A

THANKS*2 (two cents ?.. no ;p)

EDIT: for chipset that means uninstall first then reinstall full new version ?

 

[AsRock]

I've never had that happen. Since you mentioned RDR, I'll have a look once I finish Kingdom Come. Is there any game where it happens faster, and ReBAR is definitely the cause?


5000 MHz all core is really nice! That's what my 7700X can do at stock.

As far as i know it is, turn it on the game crashes, turning it off in the AMD software no crashes, this does not include the much more common crashes which is the #FFFFFFFFF error or GRX_Err which the later can be coursed by under volting.

 

[Mussels]

BETA doesn't mean bad or dangerous, it just means they aren't doing their full suite of testing on an old board.

As an example, if they updated AGESA content from AMD without altering anything else there is no reason to suspect any problems could crop up, other than universal bugs within AGESA (Like the known PBO EDC limits for overclocking)

As for that board, it's basically AGESA changes

1.70 is agesa 1.2.0.3A, which is older and had the known issues with TPM stuttering (fixable by not using TPM, but still)
1.80A is 1.2.0.7
1.80F is 1.2.0.A which isn't documented anywhere
According to this post, it may help with issues with waking from sleep and USB - the system didnt support S3 sleep states prior to that BIOS, apparently.

Only the 3150G is added after 1.70, so you don't need the newer BIOS, but the security fixes and TPM fix may be beneficial.


There is websites that let you compare the files, but i've forgotten its damned name - i'll try and find that out and compare them and get back to you

Ryzen SMU Checker (kolos.github.io)


1.80A

1.80F

Picasso, Renoir and Cezanne CPU's got small AGESA updates.
Since the 5700g is Cezanne, it's definitely got updates included - combined with the USB and sleep fixes mentioned, i'd say it's worth updating to.

 

[tabascosauz]

BETA doesn't mean bad or dangerous, it just means they aren't doing their full suite of testing on an old board.

As an example, if they updated AGESA content from AMD without altering anything else there is no reason to suspect any problems could crop up, other than universal bugs within AGESA (Like the known PBO EDC limits for overclocking)

As for that board, it's basically AGESA changes

1.70 is agesa 1.2.0.3A, which is older and had the known issues with TPM stuttering (fixable by not using TPM, but still)
1.80A is 1.2.0.7
1.80F is 1.2.0.A which isn't documented anywhere
According to this post, it may help with issues with waking from sleep and USB - the system didnt support S3 sleep states prior to that BIOS, apparently.

Only the 3150G is added after 1.70, so you don't need the newer BIOS, but the security fixes and TPM fix may be beneficial.


There is websites that let you compare the files, but i've forgotten its damned name - i'll try and find that out and compare them and get back to you

SMU checker

1.2.0.A is the new security minded AGESA after 1208, doubt there's any performance improvements

 

[Klemc]

SMU checker

1.2.0.A is the new security minded AGESA after 1208, doubt there's any performance improvements

Yes, i found a page yesterday, it lists a lot of letters/numbers refrencing security thingies...

https://www.amd.com/en/resources/product-security.html#security

@Mussels, thanks a lot.​

So, i shouldn't disable fTPM ?.. in bios !

 

[Mussels]

SMU checker

1.2.0.A is the new security minded AGESA after 1208, doubt there's any performance improvements

The latest 'final' which is only on 1.2.0.3A - since that's prior to the C fix for TPM, there very well could be a performance benefit if that issue gets triggered on the system

The betas skip quite a few updates, probably forced to by AMD for the security updates

 

[tabascosauz]

The latest 'final' which is only on 1.2.0.3A - since that's prior to the C fix for TPM, there very well could be a performance benefit if that issue gets triggered on the system

The betas skip quite a few updates, probably forced to by AMD for the security updates

In my testing most CPUs still like some variation of 1203 best. That includes the 7nm APUs - on some boards 1203 also brings in new features like GPU PBO and GFX Curve Optimizer. Neither is as great as the features on the new RDNA iGPUs though. Otherwise a lot of the later AGESAs between 1203 and 1206 are just nothingburgers - 1206 that introduced 5800X3D support, then 1207 finally properly fixes the fTPM problem.

The new 120A AGESA looks interesting, lots of security fixes. I would test the 5700G (expecting a regression), but there's no new BIOS for the Aorus AX yet.

So, i shouldn't disable fTPM ?.. in bios !

A few new Win 11 security features (VBS and HVCI) rely on fTPM. If you don't care about them and disabled the fTPM check via Rufus when installing, then it's fine to disable fTPM.

Pre-1207 BIOSes can suffer the fTPM stutter if it is enabled

 

[Mussels]

In my testing most CPUs still like some variation of 1203 best.

1.2.0.3c sure, but not A.
7 fully fixed TPM, but 3c has iGPU specific improvements relevant to his exact CPU - so he needs one of the betas to get those improvements added in between.

I use rufus to install windows so i remove the TPM requirement, so I've had the option to disable TPM in the BIOS on those if i wanted to avoid the issue

 

[Klemc]

On my B650E-E Asus Strix there is a jumper to Overvolt the CPU, if used with a 7800X3D, could it fry it immediatly ?.. or is it a different voltage thingie ?

Something i never understood is internal audio header, why is there and what's his purpose ??

I think i'm interested to add a Thermal Sensor, could it be a good idea, and is it a simple hardware or something that needs knowledge or perhaps can interfere with other components ..?
Because this PC i want to keep forever, i'm finished invest money.


Oh, and the Thunderbolt header, i never understood how to use it, what is it or even how it connects to anything, totally unknown thing to me.

//

Oh Oh, and that, perhaps i want it : AMD Secure frame Thermalright

 

[Athlonite]

Something i never understood is internal audio header, why is there and what's his purpose ??

It's for your front panel audio out and mic in

think i'm interested to add a Thermal Sensor, could it be a good idea

did you not get one supplied in the mobo box it's a simple 2 pin connector with a thermocouple on the other end

Oh, and the Thunderbolt header

it's sort of like USB but you can daisy chain thunderbolt things together like portable HDD's, external GPU's, monitors or USB hubs

On my B650E-E Asus Strix there is a jumper to Overvolt the CPU

Don't use it if you're using a X3D CPU it's purely for overclocking mostly used when cooling via liquid nitrogen or the like

 

[Klemc]

It's for your front panel audio out and mic in

NO need, thanks.

did you not get one supplied in the mobo box it's a simple 2 pin connector with a thermocouple on the other end

NO, only wifi/bt antenna... i use AIO Freezer II.

it's sort of like USB but you can daisy chain thunderbolt things together like portable HDD's, external GPU's, monitors or USB hubs

It accesses all interfaces ??? PCIe, USB, display port... how should i use it, i have never had a thunderbolt cable of my life...

Don't use it if you're using a X3D CPU it's purely for overclocking mostly used when cooling via liquid nitrogen or the like

I could try but i'm tired of RMA...

... new bios out 1.0.0.7a looks a good start, i will mount my PC friday, then backflash (waiting mx6 paste and Thermalright AM5 Secure Frame).

//

Plug that and put it near my CPU ,

XSPC Wire Sensor​

 

[AusWolf]

As far as i know it is, turn it on the game crashes, turning it off in the AMD software no crashes, this does not include the much more common crashes which is the #FFFFFFFFF error or GRX_Err which the later can be coursed by under volting.

I'll have a look - although, I haven't had any such (or other kind of) crashes with my current system. The last I struggled with these was with the 5700 XT, especially on the early drivers.

 

[freeagent]

Thermalright AM5 Secure Frame

I’ve been wondering how this performs, but no am5 here yet

 

[Klemc]

I’ve been wondering how this performs, but no am5 here yet

It doesn't, it's meant to not let paste spreeded everywhere around and even under CPU when cooler's pressed...
... and avdantage is, avoid AM5 jumpy lock system that truely ejects CPU when reopened LOL.

 

[AsRock]

I’ve been wondering how this performs, but no am5 here yet

about 2c difference (lower)

 

[freeagent]

It doesn't, it's meant to not let paste spreeded everywhere around and even under CPU when cooler's pressed...
... and avdantage is, avoid AM5 jumpy lock system that truely ejects CPU when reopened LOL.

Ohhh…

Booo. I thought it was like the intel one.