X670 resource

[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
octool port to linux

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


Test BIOSes:

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 1003
strixe-e 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

strix e f 1003
strix e f 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

strix e a 1003
strix e a 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

strix itx 1003

crosshair hero 1003
crosshair hero 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

crosshair gene 1003
crosshair gene 1004    (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

crosshair extreme 1003
crosshair extreme 1004  (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

creator 670 1003
creator 670 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

B650 creator 1003
B650 creator 1004 (fixed a hi dram v mode enabled bug + Fix Non-UEFI VGA card cannot display issue ) *note :"pulled support of 24/48gb ram due to request to wait for official firmware release"

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.

[xeizo]

fredepf wrote:
How can i get these Score on my rig?
i don'T find the way to have better score..

Post a Zentimings screenshot, we will see how your subtimings are

[Poloasis]

zedcrown03 wrote:
Gskill 6000mhz c30 EXPO kits.
now i am running 6400 C30--38-38-20 ,trefi 65535 aida64 read and write at 90000MB/S,latency 57ns.
the setting is on 0803bios ,if you use 0805 bios ,voltage will be low ,MC,VDDQ and VDD at 1.48v,soc auto

That's impressive!

[hiryu]

I ran memtest86 throughout the night. Less than 6 hours in, the testing completed without issue. Seems my problems are almost certainly not memory related. It's probably BIOS and/or vBIOS related.

I'm running 4x dimms of G.Skill F5-6000J3038F16G at 6000 Mhz on the Asus Crosshair x670 Extreme.

Now that I've been running it for a while... My box is a lot more stable with 0805. Occasionally, my GPU will still occasionally "fall off the bus" (but definitely less frequently), however, it hasn't once happened when not gaming, which is a huge improvement for the stability and general non-gaming usability of my box.

Still no vbios updates for my GPU.

So my feedback is that 0805 is working a lot better than even 0803 (which worked as well as 0705) at least for me. Looking forward to testing the next BIOS update!

[hiryu]

hiryu wrote:
I ran memtest86 throughout the night. Less than 6 hours in, the testing completed without issue. Seems my problems are almost certainly not memory related. It's probably BIOS and/or vBIOS related.

I'm running 4x dimms of G.Skill F5-6000J3038F16G at 6000 Mhz on the Asus Crosshair x670 Extreme.

Now that I've been running it for a while... My box is a lot more stable with 0805. Occasionally, my GPU will still occasionally "fall off the bus" (but definitely less frequently), however, it hasn't once happened when not gaming, which is a huge improvement for the stability and general non-gaming usability of my box.

Still no vbios updates for my GPU.

So my feedback is that 0805 is working a lot better than even 0803 (which worked as well as 0705) at least for me. Looking forward to testing the next BIOS update!

Now that I've had sufficient testing with the 0805 linked in this forum with my Crosshair x670 Extreme motherboard... I have some updates:

1. My GPU crashes had nothing to do with the BIOS... It was a cable issue. Bending the cable (it's a native 12-pin cable that came with my Seasonic PSU), caused one of the small pins (in the smaller 4 pin connector that communicates to the GPU how much power is available). I didn't bend the cable very far at all as I was already aware of the issues, but it pulled out the right most pin. This should limit the power of the GPU to 450w. The specs for the GPU show it going up to 480w... So maybe MSI needs to update their vBIOS to handle the situation appropriately? After a week of testing, including a lot of game benchmarks. Unigine, Superposition, and some crypto mining on the CPU... There hasn't been one issue. If I go back to the old cable and push the pin back in... It would very probably work. If anyone is curious, I can share pictures to show the difference. Sharing this information in case anyone else has a similar issue so they can know what else to check and that it may not be an early adoption BIOS issue.

2. Even with all the GPU stress testing, my box has been rock solid with 4x dimms of memory at 6 Ghz (with expo support)! Seems Hardware Unboxed is correct... You can get away with a good amount more than the official memory specs. The memory is single ranked. With dual ranked, you'll not be able to be as aggressive.

3. Installing a NVME SSD to PCIE slot adapter in the bottom x4 slot of my Crosshair x670 Extreme can cause the system not to POST with q-code 14. If I reflash the BIOS, I can get the system running again (so probably a CMOS clear would work, but I didn't try it). I can have the system on for the rest of the day without issue... Until I turn it on in the morning.

Using the same NVME adapter in the 2nd x16 slot has no problems. This isn't great because then you're stuck running your GPU at x8, and even if you don't care about that, it still limits the number of M.2 SSD's you can have. In the near future I'm going to migrate from 2 of my SSD's to a single larger one (going from 2 gen 3 SSD's to 1 Gen 4!) so this isn't a huge problem for me, but I'm reporting a bug.

4. I have a remaining issue: If I reboot my system and try to load an OS from the bootloader (grub) that's on a SATA SSD... The bootloader seems to not find that SATA SSD. If I try to do the same thing from a cold boot, it almost always works. I've seen this sort of issue with most of the Asus boards going back some years now and almost all saw the issue fixed eventually with a BIOS update. Hopefully sooner rather than later.

Note: I can boot to an OS from on a different NVME... Just not to a SATA SSD. Perhaps the BIOS assumes after a successful boot from an M.2 SSD, there's nothing of concern at boot time from a SATA SSD?

Anyway, hope this is helpful for other people troubleshooting problems.

[LiKenun]

Anyone having issues with the Crosshair X670E Extreme where the issue is getting stuck on Q-code 0D only after reboots/shutdowns? To get to the point where I can boot the OS, I have to clear the CMOS almost every time. It’s currently the only method that works 100% of the time for me. However, cleaning the CMOS means I also have to go through the initial BIOS all over again.

I’m on BIOS version 0805 released just two days ago (11/15/2022).

Hardware configuration:


[TH]Component[/TH]
[TH]Product[/TH]

Motherboard	ASUS ROG CROSSHAIR X670E EXTREME
Processor	AMD Ryzen 9 7950X
RAM	Kingston Server Premier DDR5 4800MT/s ECC UDIMM
Power supply	CORSAIR AX1600i Digital ATX Power Supply

Sometimes I can shutdown/reboot and get to the OS without having to clear the CMOS and set up BIOS all over again, but I have yet to methodically determine how to achieve that consistently. I’m in the process of doing repeated reboots and cold boots right now to see if that’s even possible.

I’m willing to try new BIOS versions while this machine is still on probation.

EDIT: These BIOS settings seem to be working for me. I am on my fourth reboot without having to clear the CMOS or touch the BIOS settings. (Not sure why the screenshots all say I am on 0705 because I flashed the 0805 update more than once so I’m pretty sure it’s not 0705 anymore.)

[Owner5566]

zedcrown03 wrote:

Another problem is when i use 0803 bios ,i set fclk at 2166mhz,mem at 6400mhz ,uclk will be 3250mhz,it 's unusual.
And,if i want fclk at 2166mhz/2200mhz to be stable,i need to misc voltage to 1.45V,SOC volatge to be 1.45V.(sometimes it's also unstable)

Not sure since which Bios or already all the time like this, but as far as I know everyone has this behaviour from UCLK depending on FCLK.
So I assume its AGESA related since not Motherboard (Vendor) / BIOS dependant.
But it gives also more performance, so to me it seems like if you have for example 6200 Memclock and 2100 FCLK then you are somehow running like 6300/3150/2100. Therefore you need less tight timings/more voltage for it to be stable.
Here is my discovery:

[xeizo]

Owner5566 wrote:
Not sure since which Bios or already all the time like this, but as far as I know everyone has this behaviour from UCLK depending on FCLK.
So I assume its AGESA related since not Motherboard (Vendor) / BIOS dependant.
But it gives also more performance, so to me it seems like if you have for example 6200 Memclock and 2100 FCLK then you are somehow running like 6300/3150/2100. Therefore you need less tight timings/more voltage for it to be stable.
Here is my discovery:

Very well summarized, thanks!

[zedcrown03]

Owner5566 wrote:
Not sure since which Bios or already all the time like this, but as far as I know everyone has this behaviour from UCLK depending on FCLK.
So I assume its AGESA related since not Motherboard (Vendor) / BIOS dependant.
But it gives also more performance, so to me it seems like if you have for example 6200 Memclock and 2100 FCLK then you are somehow running like 6300/3150/2100. Therefore you need less tight timings/more voltage for it to be stable.
Here is my discovery:

yeah! that's right!

[Tsikos]

Version 0804 tested boost over 5.5ghz now but still every setting in precision boost overdrive are broken give BSOD or strange numbers!

https://imgur.com/N9uWnm1

[Tsikos]

@Shamino please check this screenshots after the bios you ask me to test the system is solid with xmp ii also with today bios 0804 but my problem with pbo still exist here some screenshots!

https://imgur.com/oZfsjMY
https://imgur.com/18cl2fI
https://imgur.com/Do8XRhh
https://imgur.com/4gNCqF2