X670 / X870 resource thread

[Shamino]

ill use this thread to collect some new test bioses for the boards, maybe also to explain some less understood options

to disable cores ccd go here and choose ccd xx bit map down core.
each ones stand for an enabled core
best to disable from the back, ie:
110000
instead of 0011000
after selection press downcore apply changes or discard if made mistake

ocpak/octools

FAQ:
7950x not boosting pass 5.5G -> check that CStates is not disabled
Detailed Explanation on CState Boot Limiter


Test BIOSes:

Spoiler

new:
X3D OC Preset for those MB with asynch BCLK Support: (for simple slight perf boost for X3D)


DOCP/EXPO Tweaked: (for simple timings tightening)


strixe-e 1515 

strixe-f 1515 

strix e a 1515 

crosshair hero 1515 

crosshair gene 1515 

crosshair extreme 1515 

creator 670 1515

creator b650 1515

strix 650E I

strix 670 itx

for crosshair and strix e-e:

explanation of segment2 Loadline:

customize a heterogenous loadline for a dual segment workload range.

example above shows loadline=L6 when current is in range of 0~40A, and Level4 when current is above 40A.

Adds for x3d

dynamic ccd priority switch with core flex, os / driver agnostic so win10 win11 ok





Algo as follows:
If condition reached and ccd0 specified, then check current mem/cache activity > threshold and hysteresis reached, if fulfilled then switch
If condition reached and ccd1 specified, then check current mem/cache activity <=threshold and hysteresis reached,, if fulfilled then switch
Default hysteresis =4

Can combine multiple algos for ccd priority so combinations are wide

works on non x3d too but of course senseless on it. detailed explanation here.

[CL26_Master]

@ResUC Thanks for your answer!

Right now, I’ve set VDDG's to 1.05V and left VDDP at 1.0V. I’m currently running Karhu RAM Test with these settings, and it hasn’t shown any errors for 2 hours.

I recently applied some Bias settings based on @gupsterg ’s recommendations, and they seemed to make a noticeable difference.

Do you know more about these Bias settings or use these settings in your daily profile? What kind of positive or negative effects they might have?

DDR5 Robust Training Mode [Enable]
Nitro Rx Burst Length [8X]
Nitro Tx Burst Length [8X]
Nitro TxDFE Gain Bias P0 [DFE Gain Bias Step 0]
Nitro RxDFE Gain Bias [DFE Gain Bias Step 1]

Thanks again for your help!

 @X_Disturbed_X @gupsterg @ResUC  Do you think 6600 1:1 might actually need around 1.5V VDDIO?

[gupsterg]

Sweet! savour the taste! 😎

Personally VDDIO 1.5V I wouldn't have a concern with.

The Kingbank Dark kits I own have EXPO/XMP profiles, they default to VDD/VDDQ/VDDIO 1.5V under EXPO/XMP. Just be aware if you looked at AGESA rules for voltage VDDIO caps out lower, but I reckon this is based on early days of DDR5 when higher speeds weren't as attainable.

This OCN member ran voltages considered unsafe on Ryzen 1000 series long term (ie years) without issues. There are a few others I know of 😉 , but be aware this is all meddling users shares 🙂 . OC'ing is risk, but also seen peoples HW die with stock settings.

With DFE gain bias you are adjusting the bias on DFE taps.

Even though we can set x settings, there is some "dynamic" parameters going on.

There are videos and PDFs I have been wrapping my brain around, still learning 🤓 .

Video below touches on some DFE aspects, pay attention to around 7 minute mark, you will gain a small insight into what the DFE Taps do.

Spoiler

DFE Read Training [Enable] I set this to forced on, I do believe it defaults Enabled. You'll find it within Advanced > AMD CBS > UMC Common Options > DDR Options > DDR Training Options.

Nitro DFE Vref Offset Limits [Disabled], this has also aided me with stability on some HW at push memory clock, some it has not. This setting when set to disabled removes limits for DFE Vref offset, read the helpstring in BIOS for setting.

I don't know if you have played with ODT Group B, this can aid stability also. See how it goes for other test loads vs just Karhu on 6600MT/s before playing with ODT Group B.

VSOC / VDDP scale as we increase MEMCLK.

It won't be a linear rise, but a curve, as you push higher. Especially in 1:1, each step up of MEMCLK may need more of an increment then last MEMCLK.

I have only used 6400MT/s in 1:1, as the 4x 9000 series I have owned, showed signs when profiling, the voltage jumps needed for say 6000/6200/6400 meant 6600 would be unattainable. As I sold off 2x16GB kits and went nuts on 2x24GB, I went all out on 2:1 setups.

VDDG CCD & IOD I set same values, as the stock rule has been this since AM4, talking to others with greater insight this has been highlighted as way to set them.

My R7 9800X3D is the only 9000 series CPU I have that allows FCLK 2233MHz with stability for quite a bit of testing. This will POST FCLK 2200MHz at 900mV, a lot of testing can be done at that voltage, but 940mV is where it sits, I use 950mV to give "guard band". For FCLK 2233MHz 980mV is ample to do use with 6200MT/s and eCLK PBO CO CS for lengthy testing, see this old ZIP.

A R5 9600X I had, FCLK 2133MHz is all it did with VDDG CCD & IOD 900mV. It needed 920mV to POST 2200MHz, stability testing showed it required 940mV. I used 950mV to give "guard band", MEMCLK attained 6400MT/s 1:1, 8100MT/s 2:1.

Another R5 9600X did FCLK 2200MHz with 900mV, just couldn't be bothered to profile for this for that aspect and just used 950mV. A R7 9700X I have, I tried everything to get >2167MHz FCLK, but it just does not do it.

[Exostenza]

I've been running 3287 on my TUF X670E since it came out and everything has been fine except for some USB enumeration issues. I'm wondering if anyone is experiencing wonky USB with the latest stable BIOS with AGESA 1.2.0.3g or if it is jut because I either messed with some USB settings or got a cheap USB 3.2 cable - the thing is it worked before the update so it's unlikely to be a problem with the cord, I hope.

@SAFEDISK If you're able to pass on a request for a feature to tie the fan speeds in the BIOS to more than just one source that would be awesome, especially to the GPU. I know there is software to do this, and I might be forced to run it, but I prefer to have all my fan settings done in the BIOS so I don't need a program constantly probing my hardware in order to get a decent chasis / AIO fan curve when the GPU is hot but the CPU isn't. 

Thanks

[acivi1]

Able to add GPU and RAM temp as source for fans in BIOS would be perfect addition!

[Silent_Scone]

---

@Exostenza wrote:

I've been running 3287 on my TUF X670E since it came out and everything has been fine except for some USB enumeration issues. I'm wondering if anyone is experiencing wonky USB with the latest stable BIOS with AGESA 1.2.0.3g or if it is jut because I either messed with some USB settings or got a cheap USB 3.2 cable - the thing is it worked before the update so it's unlikely to be a problem with the cord, I hope.

---

What behaviour? Test at Optimised Defaults following a clear CMOS and report back, thanks

[GabcamMT]

@Silent_Scone  @Exostenza  9800X3D / ROG CROSSHAIR X870E HERO (1715) .... been having anomaly in usb enumartions mostly on reboot not cold poweron from day1 even with ClearCMOS/LoadOptimizedDefaults ....  even after removing all ARGB headers and almost all usb devices .....   I doubt it will ever get addressed at this rate on this gen/combo

[gupsterg]

@Exostenza 

Yeh more options would be great, below 20% PWM would be handy also.

Temp probe attached to T_Sensor and mounted close to GPU maybe a solution.

[Silent_Scone]

Some fans may not spin up properly below a certain threshold; the 20% is likely to accommodate for tolerance. Without asking, this would be my assumption 

[gupsterg]

Yeh that was stated many years back, but a few other manufacturers do allow lower PWM.

I guess work around would be UEFI default to PWM 20%, but user can select lower.

I had revealed fan off on C7HWIFI with UEFI mod, is there to make visible on 600 series also.

Spoiler

I recently tried some Arctic P12 Pro and as PWM wouldn't go below 20% I was stuck with higher RPM and also couldn't use fan off feature they have at 5% PWM.

Spoiler

[gupsterg]

Improved tCL, tRFCsb, tRDWR at 9600MT/s, more info in this post.