RaptorLake Resources

[Shamino]

i will use this as a collection of test bioses/tools/info targeted towards Raptorlake + z69/790

NOTE for Z690: You must update ME on your Z690 to properly support raptorlake (if you are on dual bioses then you need to do this to both bioses)

1) d/l and install ME driver
https://dlcdnets.asus.com/pub/ASUS/mb/03CHIPSET/DRV_MEI_Intel_Cons_ADL_TP_W11_64_V2229320_20220809B....

2) update ME with ME Update Tool (I know there are newer versions but its not up on Asus official site yet so >= this version is fine)
https://dlcdnets.asus.com/pub/ASUS/mb/03CHIPSET/MEUpdateTool_16.1.25.1885_T.zip

3) update to Latest BIOS of your choice

Note 2 for Z690: i have some Z690s that are already LOW in battery so you are advised to use a new Battery during this upgrade.

New test bioses

Z690 Apex

Z690 Extreme

Z690 Hero

Z690 Hero Eva

Z690 Extreme Glacial

Z690 Formula

Z690 Strix E

Z690 Strix F

Z690 Strix G

Z690 Strix A WIFI D4

Z690 Strix A WIFI

Z690 Strix I

Z690 Creator

Z690 Prime A

Z690 Tuf

Z690 Tuf D4

Z690 Tuf WIFI

Z690 Tuf D4 WIFI

---------------------------------------------------------------------------------------------

Z790 Extreme

Z790 Hero

Z790 Apex

Z790 Strix I

Z790 Strix F

Z790 Strix E

Z790 Strix A D4

Z790 Strix A

Z790 ProArt

Z790 Strix H

Z790 Prime A WIFI

Z790 Prime M Plus D4

Z790 Prime P

Z790 Prime P D4

Z790 Prime P WIFI

Z790 Prime P WIFI D4

----------------------------------------------------------------------------------------

Latest ocpak

Turbovcore
https://www.dropbox.com/s/ifsr8x0zncxcvyd/TurboV_Core_1.10.20.zip?dl=0

----------------------------------------------------------------------------------------

[crobartie]

I installed a new BIOS yesterday...
At first, I selected F5 (Default). It was the Performance Preferences [Intel Default Settings] option. I already noticed high voltages in the BIOS, but I booted into Windows for a quick test. In basic tests, the maximum voltage exceeded 1.4V, which is crazy (13600k with TUF z790 Plus Wifi). 

After changing various parameters, particularly:
Performance Preferences [Intel Default Settings]
SVID Behavior [Intel] -> [Typical Scenario]
the voltages dropped.

In my opinion, it's not a good practice to set over 1.4V for the 13600k (default F5).

I noticed that selecting Performance Preferences [Intel -> Asus Advanced OC] changes these parameters:
Realtime Memory Frequency [Disabled -> Enabled]
UnderVolt Protection [Enabled -> Disabled]
BCLK Aware Adaptive Voltage [Disabled -> Auto]
IA TDC Current Limit [Motherboard's Capability -> Auto]
GT TDC Current Limit [Motherboard's Capability -> Auto]
Enhanced TVB [Enabled -> Auto]
Overclocking TVB [Enabled -> Auto]
Are any of these key, and should they be enabled as the Asus Advanced OC profile does?

[Wesley1]

1.4 V at idle is rather high voltage, though it's not unseen.
Can you post a photo from bios with your voltage table? It is called V/F Point Offset in AI Tweaker.

Do you overclock your CPU?

Some example of V/F Point Offset from internet.

Example of V/F Point Offset

[crobartie]

Earlier, I wrote about maximum voltages, not at idle (although the idle voltages were also high). Currently, I've set almost everything as it was after the fresh BIOS update. What increases the voltages for me is the setting:

SVID Behavior [Intel's Fail Safe] instead of, for example,  SVID Behavior [Typical Scenario/ Auto]. When I use those settings, I don't have high voltages. However, with SVID Behavior set to anything other than [Intel's Fail Safe] (e.g., SVID Behavior [Auto] or SVID Behavior [Typical Scenario]) and IA CEP [Auto(Enable)], GT CEP [Auto(Enable)], SA CEP [Auto(Enable)], the CPU performance is HALF of what it normally should be(on the old BIOS was also OK).

The only setup that works normally for me (normal voltages and normal performance of the 13600k) is SVID Behavior [Auto / Typical Scenario] with CEP disabled, meaning: IA CEP [Disable], GT CEP [Disable], SA CEP [Disable].

In the old BIOS, I had both low voltages and CEP enabled (and normal single and multi-threaded performance, not HALF MT performance). Why does this BIOS perform worse and how to fix it? Is enabling CEP important? According to Buildzoid from @ActuallyHardcoreOverclocking, CEP should be enabled. Below, I’m attaching the photos you requested.

BIOS 1663BIOS 1663HWiNFO64 voltagesHWiNFO64 voltages

[Redline0]

Try SVID behavior - Worst Case Scenario, that's what I had to do to get my voltages and performance back to normal after the update on a 13700k, apparently worst case is still lower than Inte's fail safe.

---

I'm having a different issue after the update though, but I'm not 100% sure if the BIOS is what causes it. My network card only runs at 100 Mbps now and I have to enable/disable or unplug the cable a bunch of times before it switches to gigabit, even after manually changing the settings to 1Gbps. Already tried updating the drivers, didn't help, it's really annoying

[crobartie]

Maybe try this? https://www.intel.com/content/www/us/en/support/articles/000057261/ethernet-products/gigabit-etherne... 

[Redline0]

Oh I'm sorry I forgot to mention, my board is a Prime Z790-P, it uses a Realtek network adapter, not Intel

[Exist2Resist]

My understanding is that VID is not the voltage that is being pumped into the cores, that is the voltage that each core requests ( and does not necessarily mean that it gets it), the voltage being pumped into the CPU is the VCORE.

VID does not take Loadline and Vdroop into account. So looking at VID and is being pumped into the core is inaccurate.

[lastique]

> VID does not take Loadline and Vdroop into account.

VID is the voltage the CPU requests from the VRM. It does account for Vdroop in the sense that CPU will request higher voltage for higher anticipated current (i.e. load). This adjustment is done based on the AC LL value that the BIOS programs into CPU on boot. AC LL is supposed to correspond to the VRM and power plane design on the motherboard. Motherboard vendors used this value to effectively undervolt CPUs, which is where lots of instability reports are coming from.

But the VID values that you see in HWINFO are typically not the same as the VID requests actually sent to the VRM. The values HWINFO shows are compensated back to the voltage value that the CPU expects to receive, after Vdroop. This conversion is done using the DC LL value that BIOS also sets in the CPU. DC LL corresponds to the amount of Vdroop VRM allows for, i.e. the LLC mode that is programmed in the VRM. DC LL is also used for calculating the amount of power CPU consumes, which in turn affects how CPU maintains its operation below set power limits. VIDs in HWINFO will roughly match the VIDs sent by the CPU to the VRM if you configure DC LL = 0.01 in BIOS (i.e. if you tell the CPU that there is effectively zero Vdroop).

It is true that the Vcore is the actual voltage that gets applied to the CPU - at the point where it is measured. However, the caveat is that Vcore is typically less precise (e.g. in my experience with Z690 Hero, its granularity is somewhere around 9 mV) and doesn't catch voltage spikes very well (because it is actually measuring voltage rather than calculating it), while VIDs are more precise and more responsive. Here, by spikes I mean actual voltage in response to a high VID rather than a load release overshoot.

Buildzoid has a good video about it, although it's rather long:

https://www.youtube.com/watch?v=oIIj9kJHf6o

[Huigie]

Nothing wrong with 1.4V during idle or light load. You only have to worry about high voltage @ high loads.

[crobartie]

Oh, I didn't know. Under heavy load, I have below 1.2V. 
I was focusing on the maximum voltage values, without considering the type of load. A big part of this issue with Intel stems from excessively high voltage spikes that degrade the CPU and eventually lead to instability, right?