EdgyBloke
Ok I’ve been talking to EVGA and let’s clear a few things up. There are issues, but some reports on here are nonsense.

No this BIOS won’t damage your board, but it does seem to treat Vcore differently than 1.03. Now another issue is that ring is absolutely locked to 4.5GHz. After extensive testing these are the only 2 oddities I can confirm. If your downclocking it means your thermal throttling, because tbh most overclocks people show me are wildly unstable or have way too high of voltage realistically.

EVGA is already investigating the ring lock and it’s likely 2.05 will be available soon. I’ll talk to them again and see if they are able to push it out early this coming week, but idk a for sure timeline.

This is nonsense. I am NOT thermal throttling and DO NOT have this issue on v2.03 or any other prior firmware version. This assumption is inappropriate and based on nothing. I agree that some people lack experience and do not know how to overclock, but some of the people in this thread, myself included, are highly experienced and respected overclockers with the credentials to show for it. Let's not dismiss the issues this firmware has with assumptions they are driven by user error.
 
They had a similiar issue with CPU clock throttling on the Z590 Dark that was due to "VMAX Stress" being enabled on a menu they had hidden from us. I unlocked the hidden menus, exposed the setting, disabled it and the problem was solved. It was fixed on the next BIOS update. Probably something similar here. Some of the drama could be avoided by exposing more menus that are currently hidden (but commonly exposed on competitor motherboards). It is unclear why and kind of illogical that they are hiding some menus that are normally accessible... I really don't get that.
 
Thanks for taking time to visit with EVGA. The fact that we can is one of the things I love most about EVGA. Good luck having any kind of meaningful interaction with their competitors.

EdgyBloke
Idk what to tell you. You can report it, but I can’t reproduce your issue on 13th Gen unless I thermal throttle. Dropping to 800MHz is an indication of thermal throttling when you hit 100c. Outside of that check to make sure all your cables are fully seated and try using the CMOS reset button. Keep in mind even at stock settings the 13900K can thermal throttle even with good cooling.

Especially if you suddenly add 30mV of extra voltage which 2.04 seems to do at least for me.

I fixed it by flashing back to v2.03. Problem immediately solved. Running on chilled water with core temps in the 70's, so no... no thermal throttling.
 
I may unlock all of the hidden menus so I can do my own testing. Probably something simple (human error on EVGA's part) like leaving VMAX Stress enabled like what they did on the Z590 Dark.

Dear Zippytec, let me try to explain better:
When I have the 7200 A-Die kit in the motherboard with the 2.04 Bios, I can only get to boot if I don't change any BIOS settings.
If, for instance, I leave all Bios settings on default and only change the FLL Override to Mode 1, the board goes to the 53, 54, 7F loop. And only gets out if I reset the bios on the back button.
This is only happening with the 2.04 Bios and this particular kit of 7200 A-Die. If I change back to 2.03 it works OK with the same memory kit.
In my particular case, I can only use the 2.04 with the Teamgroup T-Force 6400 memory kit.
 

Kadao
Dear Zippytec, let me try to explain better:
When I have the 7200 A-Die kit in the motherboard with the 2.04 Bios, I can only get to boot if I don't change any BIOS settings.
If, for instance, I leave all Bios settings on default and only change the FLL Override to Mode 1, the board goes to the 53, 54, 7F loop. And only gets out if I reset the bios on the back button.
This is only happening with the 2.04 Bios and this particular kit of 7200 A-Die. If I change back to 2.03 it works OK with the same memory kit.
In my particular case, I can only use the 2.04 with the Teamgroup T-Force 6400 memory kit.
 

has this kit ever booted past the bios? what other settings did you change? i didn't see much impact at all with FLL override in my testing on 2.02 either. can you try just setting xmp and trying to post?

Actually where do you find the spread spectrum setting? I’ve been trying to find it so I can mirror your settings.

XMP Voltage is 1.45V. I put the 7200 kit in the motherboard and it immediately goes in the loop. Reset the bios. Enable extreme voltage mode. Restart, it is working, go back to the bios and get the message "memory has changed", enable XMP and get stuck again on the 53, 54, 7F loop. 

Ok. I will warn you that having hidden settings enabled is something EVGA has likely not tested for stability, so I suggest not enabling any of those and checking that as well.

applying ring clock 50. but 45 clock.. error

Mr. Fox
 
Now the core downclocking issue is not happening like it was. I am not sure what changed, but I am applying "ratio limit" for my fixed core clocks. I will change it to "per core" and see if the downclocking returns.

Strange. It did not. So, going back to v2.03 again until this mess gets fixed.

TheAffxct
I did some more testing and I guess 7800-8000 is easier to attempt now. The BIOS started behaving better after doing a flash via the flashback feature. My AIO_PUMP header still can’t dynamically control RPM which still does not make any sense to me.

I dunno what to think about temps and voltage. R23 requires 1.2Vmin for 54/42/45 with my 13700K and maxes in the low-mid 80s. LinX with the latest binary passed at LDA 28000/6GB at the same voltage and LLC at 1.19Vmin, but hit like 93 max temp. Needless to say, I only looped it X25 because I would rather not degrade my chip.

8000 definitely isn’t happening, but it’s training consistently, and 7800 looks like I’m a few adjustments away from HCI stability. Something I don’t understand is that VCCSA Auto and 1.25-1.3 TX seems to be the play now?

TX? What setting is that?

sfusco
I installed 2.04 on my system while running EVGA RTX 3090 ti FTW3 Ultra Gaming GPU.  No issues.  I was able to snag a Zotac RTX 4090 from Microcenter for MSRP today - after install, getting code loop of 53, 54, 7F that repeats indefinitely.  Tried my buddies MSI RTX 4080 and that worked no issue in my system (EVGA z690 Classified).  I tried flashing back to 2.03 and same issue.  Flashed back up to 2.04 for the time being.  I installed the Zotac on my buddies system and it worked no issue... so then I booted up with it in my system, unplugged.  While in Windows I plugged it in (yea not a great idea but i had to try it), it turned on and i installed drivers no issue, even ran benchmarks.  Restarting PC even worked ok.  However, when shutting off and turning back on, it would go into the same memory loop cycle of aforementioned codes.  Any suggestions?

Update the VBIOS with Nvidia's tool on this page: https://nvidia.custhelp.com/app/answers/detail/a_id/5411/~/nvidia-gpu-uefi-firmware-update-tool
 
It's designed to work with all 4090's and 4080's regardless of brand.

Mr. Fox

TheAffxct
I did some more testing and I guess 7800-8000 is easier to attempt now. The BIOS started behaving better after doing a flash via the flashback feature. My AIO_PUMP header still can’t dynamically control RPM which still does not make any sense to me.

I dunno what to think about temps and voltage. R23 requires 1.2Vmin for 54/42/45 with my 13700K and maxes in the low-mid 80s. LinX with the latest binary passed at LDA 28000/6GB at the same voltage and LLC at 1.19Vmin, but hit like 93 max temp. Needless to say, I only looped it X25 because I would rather not degrade my chip.

8000 definitely isn’t happening, but it’s training consistently, and 7800 looks like I’m a few adjustments away from HCI stability. Something I don’t understand is that VCCSA Auto and 1.25-1.3 TX seems to be the play now?

TX? What setting is that?

TX VDDQ?

Mr. Fox

TheAffxct

Mr. Fox

TheAffxct
I did some more testing and I guess 7800-8000 is easier to attempt now. The BIOS started behaving better after doing a flash via the flashback feature. My AIO_PUMP header still can’t dynamically control RPM which still does not make any sense to me.

I dunno what to think about temps and voltage. R23 requires 1.2Vmin for 54/42/45 with my 13700K and maxes in the low-mid 80s. LinX with the latest binary passed at LDA 28000/6GB at the same voltage and LLC at 1.19Vmin, but hit like 93 max temp. Needless to say, I only looped it X25 because I would rather not degrade my chip.

8000 definitely isn’t happening, but it’s training consistently, and 7800 looks like I’m a few adjustments away from HCI stability. Something I don’t understand is that VCCSA Auto and 1.25-1.3 TX seems to be the play now?

TX? What setting is that?

TX VDDQ?

I do not have any BIOS setting named TX VDDQ. I have CPU VDDQ and Memory VDDQ options in my EVGA BIOS menus. It is bad enough that each computer OEM uses different terminology rather than speaking the same language and having a standardize vernacular, it would be great for a lot of people reading information online if we refer to what we find in the EVGA BIOS as it is found. 
 
Yes, you are correct. I had my VSA (or VCCSA as it is "not" called in the BIOS) set manually and could no longer boot and had to set it to "auto" to boot. I have it set manually again at 1.350V and it is booting, but 100mV more VSA than I needed before. I don't like "Auto" because I don't see a place in the EVGA BIOS that actually tells me how much is being applied. Unless I set it myself, I don't know what it is doing on its own, and that kind of sucks.

Auto on previous BIOS implementation for RPL was up to 1.45V. Since 2.04, Auto sets me to 1.3V and that’s enough to reliably power the uncore for D5-7800/8000. The naming conventions are kinda simple once you get used to them.

ASRock:EVGA:ASUS:MSI:GIGABYTE (IVR VDDQ Transmitter/TX)
VDD_IMC:CPU CDDQ:IVR Transmitter VDDQ:CPU VDDQ:DRAM Vdd/Vddq Voltage

Worst award goes to GIGABYTE, second worst goes to ASRock but I can see what they were going for, technically the IVR Transmitter is the IMC rail.

ASRock:EVGA:ASUS:MSI:GIGABYTE (VDD2/Drive rail)
VDD_CPU:VDD2:IMC Voltage/VDD:CPU VDD2:VDD2

Worst award goes to ASUS; like what the hell? VDD2 isn’t even an internal rail. Way to confuse your customers. It still cracks me up when people try to minimise VDD2 thinking it’s going to conserve the longevity of their IMC. Otherwise, ASRock here have the right idea, but they’re way too technical. The VDD2 voltage does hit the CPU, but they should’ve just went for VDD_PHY or something by that point.

Dear all,
 
I just noticed now that when I'm running BIOS 2.04 with both V-Color Manta Xprism DDR5 7200 Mhz, the motherboard only detects one stick. If I have only one stick inserted I get the 53, 54, 7F loop, it doesn't matter which one because I've tried both. With only one stick it gets in the loop as soon as I power it up. With both sticks, doesn't matter the order, BIOS only detects one stick.
 
If I go back to 2.03, it is working normally with both sticks @ 7200 Mhz.
 
 

The only 2 issues I can confirm and reproduce as of now is the ring lock and the 30mV extra supplied voltage. 2.04 has specific improvements for A-die. If you are on M-die there isn’t much point to upgrade.