ORIGINAL:
Hey all. I'm trying to set a static overclock for my 13700K on a Z690 Kingpin. Something like 5.5 or 5.4GHZ boots up just fine. Problem is, when I start up a stress test (in this case, P95 Small FFT) it automatically downclocks to 5.3GHZ. No matter what. I've checked all the settings I'm aware that could be causing this. I've tested Manual - PerCore and Manual - RatioLimit. I've tested with EIST and Turbo Mode on/off. Yes, AVX offsets are zero. Windows shouldn't be downclocking it. All the other power saving options are disabled. Static voltage. Any ideas? (Edit: I should add, this is BIOS 2.06. Though I skimmed through the updates page and didn't find any fixes related to this in future versions.)

Thanks
 
(EDIT) READ ME:
Exhaustive summary of all of my comments below + extra information.
- I've tried both BIOS 2.06 and 2.13 (latest)
- Yes, AVX offsets are OFF
- Set power limit to max (4096)
- Temperatures are good (maximum on CPU is 75)
- VRM temperatures are good (maximum is ~40)
- I've tried changing VRM switching frequency (500, 800, 1000)
- I've tried adding voltage (1.3 -> 1.35) and changing LLC (-50%, -75% vDroop)
- I've tried downclocking ring for stability (4.8 -> 4.5)
- Hyperthreading, C-states, e-cores, virtualization, TVB, adaptive boost, legacy game compatibility mode all OFF (I'm aware BIOS states that TVB needs to be OFF for overclocking)
- Tested with EIST, Turbo Mode, Adaptive Boost Technology ON/OFF (I'm aware BIOS states EIST/Turbo Mode is needed for overclocking)
- Tested both available microcode options
- TjMAX is 100c (CPU never reaches this)
- Tried both options of CPU multiplier control
- HWiNFO64 only says "IA: Max Turbo Limit"
- Using XTU doesn't change the behavior
- Tried downclocking RAM (to "automatic" setting)
- Tried repasting CPU and checking mounting pressure of CPU contact frame (worth noting I am using a contact frame)
- I've tried even higher clocks. Both 5.4 and 5.5 downclock to 5.3
- It seems this issue has been posted about before. Won't let me post links here, but google something like "EVGA AVX Limit Raptor Lake"

What EVGA suggested: Check for issues in thermal paste application and try a different PSU.
I've already checked thermal paste application. Which leaves, to the best of my knowledge, these explanations:
- Power delivery issues (will test this within a week)
- CPU hardware/firmware/whatever issue
- Motherboard hardware/bios issue
- I've read of similar issues (see above) that say there's a sort of AVX limit on Raptor Lake. This is, apparently, at 5.3GHZ, which is... convenient. If you'd like to test this, try seeing how high of an AVX clock you can get. Apparently, these are the maximums for Raptor Lake: (i7K - 5.3; i9K - 5.5; i9KS - 5.6).

My temps are good. 70C max. BIOS 2.06 doesn't offer power limits so doubt it's that. I'll try again later with latest BIOS.

Alright. Tested with BIOS 2.13, max power limit. Still downclocking unfortunately. :(

What are your settings in the Advanced-tab -> CPU Configuration?
 
I'm specifically curious about Adaptive Boost Tech, TVB, and Enhanced TVB.

Kr_cK
Late response, sorry.

Advanced -> CPU Settings:
The same behavior persists with EIST, Turbo Mode, and Adaptive Boost Technology either ON/OFF. (I am aware the BIOS notes that EIST and Turbo Mode is needed for overclocking)

OFF/DISABLED: C-States, all virtualization options, hyper-threading, legacy game compatibility mode, adaptive boost technology, TVB and all of its options (since this is a static overclock)

E cores are disabled. Power limit 4096. Microcode 0x105 (but I've tested both). TjMax 100. (My CPU never reaches this)

1.3V vCore, 1KHZ PWM frequency, and default LLC are the only voltage settings that I've changed. Ring is at 48, AVX offsets off, and I've tried both manual options of CPU Multiplier Control at 5.4GHZ.

Thanks

(EDIT: I just tried TVB on with downbins at zero. It's still downclocking.)
(EDIT 2: On HWiNFO64, the only performance limit reason listed is "IA: Max Turbo Limit". This shows up as soon as P95 is started and goes away as soon as I stop it. I just can't figure out why it's doing that during stress test only.)
(EDIT 3: VRMs don't seem to be overheating either.)

Would if you use the Intel XTU to set the frequency higher?
 
VRMs should handle 400W+ no problem for KP

Okay. I repasted the CPU (per EVGA's suggestion) and double checked the mounting pressure of my contact frame. Still downclocking. After I make sure it's not a power delivery issue, I guess I'll have to chalk it up to a hardware issue with either the CPU or motherboard. I'm going to edit in all the information I've gathered to my first post in hopes it makes it easier for other people to digest.