X870E Nova D4 on restart/cold boot

Hello, i am troubleshooting freshly built PC, seeking advice and suggestions how to proceed with debugging the issue.

AsRock X870E Nova Wifi occasionally throwing d4 error on powerup / restarts, without clear pattern. Waking from sleep/standby without error.

Hello, so i've been trying to figure it out about a week now - EXPO has been in use twice - just as a one more step towards making it work - tp be clear, BIOS is all at default settings, or in other words, no matter the settings, error creeps in occasioanlly.

Let me list whole system spec first:

Kingston 2x32GB RAM (using default Jedec profile, 4800 MHz)
Asus Prime Radeon 9070XT 16GB OC
Corsair HX1200i
Samsung 9100 PRO NVME as system drive, in PCIE-gen5 slot.

3 more drives, unused, one of them formatted for backups first time today:
- WD Green 2TB NVME (unintialized, unformatted)
- Crucial C100 1TB NVME (unintialized, unformatted)
- WD Blue 500 GB NVME (initialized and formatted on April 22nd '26).
- Win 11 PRO (reinstalled once)

All in a Phanteks Enthoo Primo, with bunch of fans for good airflow - ambient temp 24°C, CPU idling ~44°C-45°C.

This is to be my new dev system (working on large codebase, requiring fast builds and sufficient amount of RAM).
I have been using it very very lightly, as setting up local repo and tooling requires time - i have trusty old 5700x still in use, new one is on another desk.

Once i noticed d4 error first time, figured it occurs on POST only and have restarted, turned it off, left it unplugged, left it in standby for many many times in order to see if there is a pattern - no obvious pattern, sometimes it fires up, sometimes no go. I do have useful info as i have gone over three different BIOS versions.

I was always using AsRock driver updated and AppShop to pull correct drivers.

Mainboard originally came with BIOS v3.40. (i have bought everything back in December 2025, but due to various circumstances only found the time last week to assemble everything).

So i have tried following BIOS versions:

- v3.40 (had it originally)
- v4.10 (made d4 error popup MUCH worse)
- v3.50 downgrade, and what i am using now - error creeps in way less often

I'll list now what i have tried so far. BIOS settings are always assumed to be Default/Auto unless stated otherwise. I'll also note which BIOS versions where used for steps listed.

- (BIOS v4.10) reseating RAM (always two sticks, have yet to try using single stick or A1 / B1 slots)

- (BIOS v4.10) - tried to turn on EXPO, in order to see if that might actually solve issue, with manual SoC voltage (tested values: 1.05V, 1.1V, 1.15V, didn't change behavior, made d4 happen more often, with VCore Auto setting going wild in Windows (up to 1.35V)

- (BIOS v3.40) Limit TDP to 105W setting - no observable difference to d4 occurence

- Upgrade BIOS to v4.10 - first boot after upgrade, training phase (code 15) took quite a long time, but it booted

- (BIOS v4.10) - Default/Auto settings - d4 error occurs quite often

- (BIOS v4.10) Manual SoC voltage to 1.10V, limit TjMax to 85°, LLC 2 (default 3) - made d4 occurence worse, especially lower voltages, but decided manual SoC

- (BIOS v4.10) Power Supply Idle Control set to Typical Current Idle - doesn't seem to have effect on d4 occurence

- (BIOS v4.10) Memory Context Restore - disabled, Fast Boot - disabled - this made it occur less often, but still it occured.

So all of the above, it would randomly occur on restart or cold boot, but often enough to be an obstacle - when noticed it first and informed myself about it - i started to restart or power it off / turn back on quite often during past days in order to try figure out if there is any pattern.

====================================================================
Now onto final thing, which made it occur way way less:

- downgrade BIOS to v3.50 - first boot after upgrade, training phase (code 15) took much much less time, compared to first training v4.10 BIOS took)

- (BIOS v3.50) - Default/Auto settings - d4 error occurrence rare, compared to v4.10

- (BIOS v3.50) - Memory Context Restore - disabled, Fast Boot - disabled - d4 error occurence rare.

So net total is with BIOS v3.50 i have way less occurences of d4 error, almost to the point i thought it is resolved.
When it first creeped in, i went into BIOS and disabled Memory Context Restore and Fast Boot - believing that would finally resolve the issue, but it did creep back in.

Finally, when error occurs, what i have found is: system will boot almost always next time when pressing Reset key.
Cycling power when error occurs, system usually won't boot. Both of these apply to BIOS v3.40 and especially to v4.10, because i had most errors happen on those two - with v4.10 being the one where i had by far a ton of them.

Xaltar wrote: It may be worth testing a single stick in slot A2 and see if the issue persists. If not, then try using slots A1 and A2 (single channel) and see if the issue arises again. If it does then try the B channel in the same way. If the issue only occurs with 2 RAM modules installed then it could point to a marginal/weak IMC on your CPU. If the issue persists with a single module, try the other one and verify that you don't have a bad/intermittent RAM module.

Thanks alot for prompt replies, @Xaltar! I will get to memory swapping, probably not today, but will let you know how that goes.

Xaltar wrote: Those are both known to exacerbate RAM compatibility/training issues. The fact that the issue is more frequent on 4.10 tells me that it is definitely either your RAM not playing nice with the system occasionally or the IMC (memory controller) in the CPU. 4.10 introduced much more conservative voltages and sub timings, if the CPU/RAM/IMC need more voltage for stable operation then that would explain the increase in frequency of the D4 error.

That is good info and sheds light on why newer versions in this case yields more error occurrences.

Today i've restarted it about 10 times, without issue.

Most of the time it was just playing some music, connected via Toslink optical into receiver.

====
It's currently running Furmark GPU stress test, simulatanously running Furmark "CPU Burner", 24 treads, roughly 40 minutes in.

CPU temperature is rock-solid stable at 76°

PSU draws about 450W from outlet (using Corsair iCue tool, PSU has USB connection to mainboard)
12V (reading by HWInfo) - very stable, narrow regulation, min=12.038, max=12.096, avg=12.076

VCore is almost locked in at 1.112V
SoC 1.085


I know this ain't good test for memory since it's not doing much, but had to do this too, this would be the first time it's somewhat loaded.

====

Xaltar wrote: Does the issue only occur when powering on from the system being off or does it occur with restarts at all?

It did occur with restarts too - much less often then pure power-up.

On BIOS v3.50 only once on restart so far, but sample is too low to draw conclusion with any sort of confidence, just compared to v4.10 less often.

It never occured on waking up from standby in any BIOS version of setting combinations.

It occurs on power-up, being unplugged before or not, makes no observable difference.

Resolution is usually always the same, press "Reset" button and it gets thru after that.

Sometimes when error would've pop-up, i would power it off, unplugging USB drives, after which it would go thru - but i'm certain this is just a coincidence.

====

Xaltar wrote: A cold boot D4 error could also be caused by any number of things from a weak IMC to RAM to a power supply with marginal 12v power that takes a while to "warm up" and stabilize.

I can also measure 12V rails externally - will see how i can approach this meaningfully. (Suppose given it's 1200W PSU it probably has several rails, somehow doubt it's designed with allowing full amperage thru single 12V rail.)

Short disclaimer: testing i did i didn't do in "scientific" way, but can tell frequency of events going up or down without exact noting down of how many restarts or powerups it had gone thru.

As a matter of fact, my current machine (old B450 MSI board, 5700X) is usually powered on, up and running for weeks, and connected to UPS - so in a way, i could live with PC that is not so easy to startup, but still everything being brand new and top-tier it's a bit unamusing to say the least, like an old car with carburettor on cold winter morning hehe.

Anyway, once more cheers for the help

A bit of a random thought here, D4 is stated to be "PCIe initialization or RAM related
issue" according to ASRock's debug code description. Have you tried without the
GPU installed? I don't see your CPU listed but so long as it isn't an "F" variant
it should have an iGPU.

The D4 issue has been tickling at my brain since you posted about it, I was sure
I had seen another report about it before. I finally remembered. Another user
had a very similar issue, infrequent D4 codes on POST that, in their case escalated
and got worse over time. Eventually after replacing every component in their
system (every single component including PSU and GPU) the issue still persisted.
In the end it turned out to be their PCIe riser cable that came with their case....

With the weight of modern GPUs similar issues could arise simply from the weight
of the card causing it to make poor contact in the slot due to sagging.

I have a feeling this isn't the issue in your case given the difference in behavior
based on BIOS version but it's worth mentioning.