Crosshair VIII 2501's for testing

[Shamino]

still based on Patch C, should resolve some dram issues seen

https://www.dropbox.com/s/ojf8g8g1h7ot18j/ROG-CROSSHAIR-VIII-IMPACT-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/a4w0dghqe0qxvgy/ROG-CROSSHAIR-VIII-HERO-WIFI-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/1c3p11iloj3orbj/ROG-CROSSHAIR-VIII-HERO-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/o31pqe91pe15zm6/ROG-CROSSHAIR-VIII-FORMULA-ASUS-2701.rar?dl=0
https://www.dropbox.com/s/24b0vppywj5r91b/ROG-CROSSHAIR-VIII-DARK-HERO-ASUS-2701.rar?dl=0

Edit updated links to 2701
bios 2701, names in links
fixes some bugs and raises settable fmax offset.
this is still based on the same AGESA version as 250x so dont expect any difference in Fclockiness against that version.

[78Staff]

Royal 3600 C14's (2x16) arrived this week, so far so good. They are 1.45v vs 1.35 at DCOP, I guess the tighter timings require more voltage? Probably wasting them somewhat as I am running pretty much stock DCOP and basic 13% AI3 based OC. But I may tweak the settings somewhat down the road.

Loaded up bios 3003 (from 2702), running for almost 2 days now with no issues - but again running I'm 3900x and DCOP settings, nothing fancy. I've given up on 4 dimms for now. IF it all gets sorted out I may try to run 64GB again with 4x16's (for content - LR and PS mainly, but also run some VM's as well), but otherwise any cpu/gpu upgrades seems to be a ways off anyway based on lack of availability of the 5950x and 6x00 series

[RedSector73]

I am using the new 3003 BIOS with 3900x and not had a single bad BIOS except one beta that was never actually released. Your issue DrSlump76 is not the BIOS.

[the_patchelor]

running/testing 3003 on my impact with 32GB B-Die and 5950x. Same issues as with 2402. FLCK >1900 is possible BUT:

- 1.8V PLL voltage has to be set now manually to 2.25 for FCLK2000, auto will set it to regular 1.8v but then no boot or slideshow/slow boot
- if 1.8V PLL voltage is below the "required" minimum (e.g. 2.1v) the CPU performance is cuted down to 30-50% (CB R20 will give 3000-4000 points and Aida memory latency is 65-70ns for 4000CL15)...once I raise the PLL voltage speed get "normal" but still latency could be better...
- Vsoc also has be be increased (right now 1.25v, below the performance will be reduced too
- again 100x of WHEAs (CPU Bus/Interconnect errors)

OR does it mean, "hey, your CPU is nice but not capable of running higher FCLK"

...may I just should buy a cheap b550 Board and enjoy a higher FCLK :confused:

[Peter_gr]

the_patchelor wrote:
running/testing 3003 on my impact with 32GB B-Die and 5950x. Same issues as with 2402. FLCK >1900 is possible BUT:

- 1.8V PLL voltage has to be set now manually to 2.25 for FCLK2000, auto will set it to regular 1.8v but then no boot or slideshow/slow boot
- if 1.8V PLL voltage is below the "required" minimum (e.g. 2.1v) the CPU performance is cuted down to 30-50% (CB R20 will give 3000-4000 points and Aida memory latency is 65-70ns for 4000CL15)...once I raise the PLL voltage speed get "normal" but still latency could be better...
- Vsoc also has be be increased (right now 1.25v, below the performance will be reduced too
- again 100x of WHEAs (CPU Bus/Interconnect errors)

OR does it mean, "hey, your CPU is nice but not capable of running higher FCLK"

...may I just should buy a cheap b550 Board and enjoy a higher FCLK :confused:

Hi mate!

I am using 3003 bios, i have ch viii hero with 5900x and g.skill F4-3200C14D-16GTZR (b-Die).
When i use docp profile (for 3200cl14) i have random restarts.
When i overclock my memory to 3800mhz cl15 or cl16 (testing) i have zero restarts... The voltage i use is 1.365v (1.35v default).
FLCK is 1900MHZ. In AIDA the latency is between 56,4ns - 57,5ns
Cinebench R20 the score is SingleCore 594pts and Multi 7829pts. Is it a bit low?
When i am tring FLCK 2000mhz no boot at all 😞

[1ah1ksa]

I test 3003 and i am half way of fixing my 1900 fclk issue
at first i boot with 1900fclk and 3733mhz and works before 3003 bios i cant do 1900fclk alone
Second reboot i did 1900fclk and 3800mhz and it works without issue and i did some test and i got 52ns latency
I was very happy

After 10 hours of playing i did restart so i can enable pbo after enabling pbo and restart it wont boot

I think there is a bug with pbo

Can any help me where can i find the cold boot dram voltage

Thank you

[evilhf]

1ah1ksa wrote:
I test 3003 and i am half way of fixing my 1900 fclk issue
at first i boot with 1900fclk and 3733mhz and works before 3003 bios i cant do 1900fclk alone
Second reboot i did 1900fclk and 3800mhz and it works without issue and i did some test and i got 52ns latency
I was very happy

After 10 hours of playing i did restart so i can enable pbo after enabling pbo and restart it wont boot

I think there is a bug with pbo

Can any help me where can i find the cold boot dram voltage

Thank you

although with only 2 sticks I get 2000 fclk / 4000 in memory only in 2402 bios, with 4 sticks I'm limited to 1900/3800 in any bios (currently using 3003 bios).
The performance is very good and the last BIOS has increased the L3 cache rates considerably.
Dream of using 2000/4000 with 4 sticks 😎

[DrSlump76]

RedSector73 wrote:
I am using the new 3003 BIOS with 3900x and not had a single bad BIOS except one beta that was never actually released. Your issue DrSlump76 is not the BIOS.

Hi RedSector73,
according to asus customer care my problem is caused by my memory kit that's not included in qvl.
My system seems to be stable with all relased bioses (except the 2004 that has been removed from downloadable bioses list, where i had the same issues i'm having with the new bioses) until bios version 2206 that's the last stable bios.
With this bios i have no problem raising my memory kit to 4000Mhz or 3733Mhz with fclk 1866Mhz that's the max stable fclk at default voltage for my system (i haven't tried too much for lack of time).
With all new bioeses, starting from 2311, if i try to rise up fclk beyond lets say 1600Mhz (but i'm not totally sure) i get a lot of "WHEA"s in HWinfo and random system reboot when in idle.

I would really appreciate your help to understand what kind of problem could cause instability with new bioses.

Thankyou 🙂

[RedSector73]

@DrSlump76 can you fill your system specifications, otherwise I have to pull out my trusty magic eight-ball and guess and .....

WHEA errors may/may-not be a RAM issue. I have a clue for this one, disable c-states has been the resounding sucess at stopping these errors. Which suggests that for some reason CPU power is going too low when idle. Disabling c-states (power saving feature) seems to be a temp fix for WHEA problems and will see slightly higher idle temp as well.

Found the post of changes made for WHEA errors, as follows:-

What seems to have resolved this issue for me is:

1. Updating to bios 3003
2. Disable CPU C-States & DF Cstates under Advanced-> AMD CBS
3. Set "Power Supply Idle Control" option from AUTO to "Typical Idle Current";
4. Enable CPPC & CPPC Preferred cores (not sure how much this affects stability);

I can replicate the hard reboots consistently, by enabling/defaulting the above settings.

To be safe I also did the following in windows:

5. Disable all power saving features in windows and set to max performance;
6. Update all windows drivers.
7. Run "sfc /scannow" followed by "Dism /Online /Cleanup-Image /RestoreHealth" to ensure my windows installation was not corrupted.

I've never personally had the WHEA error issue but since YMMV this is a suggested temp fix.

[DrSlump76]

Sorry RedSector73, you're right.
This is my config:
https://valid.x86.fr/912ctc

So, if i correctly understood, yout system is stable when you enable the "idle states " fixes.
What's the latest bios you can run stable on your system without disabling or limiting power saving features?

Thanks

RedSector73 wrote:
@DrSlump76 can you fill your system specifications, otherwise I have to pull out my trusty magic eight-ball and guess and .....

WHEA errors may/may-not be a RAM issue. I have a clue for this one, disable c-states has been the resounding sucess at stopping these errors. Which suggests that for some reason CPU power is going too low when idle. Disabling c-states (power saving feature) seems to be a temp fix for WHEA problems and will see slightly higher idle temp as well.

Found the post of changes made for WHEA errors, as follows:-

What seems to have resolved this issue for me is:

1. Updating to bios 3003
2. Disable CPU C-States & DF Cstates under Advanced-> AMD CBS
3. Set "Power Supply Idle Control" option from AUTO to "Typical Idle Current";
4. Enable CPPC & CPPC Preferred cores (not sure how much this affects stability);

I can replicate the hard reboots consistently, by enabling/defaulting the above settings.

To be safe I also did the following in windows:

5. Disable all power saving features in windows and set to max performance;
6. Update all windows drivers.
7. Run "sfc /scannow" followed by "Dism /Online /Cleanup-Image /RestoreHealth" to ensure my windows installation was not corrupted.

I've never personally had the WHEA error issue but since YMMV this is a suggested temp fix.

[WhoSaidThat]

DrSlump76 wrote:
Sorry RedSector73, you're right.
This is my config:
https://valid.x86.fr/912ctc

So, if i correctly understood, yout system is stable when you enable the "idle states " fixes.
What's the latest bios you can run stable on your system without disabling or limiting power saving features?

Thanks

Just a heads up that in "2. Disable CPU C-States & DF Cstates under Advanced-> AMD CBS" you might want to try disabling C-states first and see if that fixes your issue/s as the cost is a relatively minor impact of around 7W to 8W on your fabric power consumption at idle. Disabling Global C-State Control will adversely affect the performance of your CPU.