5950x won't post with DOCP enabled

Just today got around to installing a 5950x that’ll be tested for performance and in all likelihood sold when all is said and done.

What is frustrating however is that it will not post with the same 4x16GB 3200MHz RAM that I’ve been using with my 3950x if I enable the DOCP profile. It won’t crash out or boot loop, it’ll just sit there like some gormless mute giving me q code 15.

I’m using this CPU in a Crosshair VIII motherboard with the latest BIOS update. I don’t think this particular RAM kit is on Asus’ QVL but G.Skill will vouch for its compatibility.

Any ideas how to get around this issue? Manually setting the timings is an option but it’d be nice to have an apples to apples comparison with this to the 3950x I’ve been using without thermal paste for the past year and change.

Nudge the voltage up .05v and try that ?

Also reorder the dimms so they are in opposite spots of where they are now

Could that really make a difference when all slots are populated with dimms from the same kit?

yep

I had a similar issue, also on Ryzen but not sure what model, which I dealt with by adjusting the memory frequency, not all the sub-timings. As I recall, one notch either below or above the nominal frequency (3133 or 3266 MHz) worked in my system, even though 3200 did not.

It might make sense to enable DOCP with a low-ish frequency, even the JEDEC freq, and work your way up.

I realize you said you’d rather not do this, so just FYI.

Well I tried.
1.5V on the RAM, 1.1V on the SoC, no dice.
Not only does this CPU refuse to post with the basic bitch xmp enabled that’s been validated by the vendor but it’s not even stable in regular operation with the default JEDEC speeds and timings. I’ve had several blue screens in just a few days.

To be fair my system is put under an abusive workload with two GPU’s mining 24/7 and several VM’s being accessed from terminals throughout the house during the day — but unless I start doing really silly shit my 3950x will handle it all just fine. This 5950x I’ve got is just not behaving as it should.
Unless there’s some simple fix I’ve overlooked this is going back with an RMA.

Wait what gpus? Pice4? If so try running them in pcie3 mode.

Could be mobo or cpu either one if hw issue

I’ll try pulling stuff out to see if it’ll post with one GPU, no nvme and only two dimms. Beyond that I’m out of ideas.

One dimm in the farthest slot?

I solved this issue by moving from the A1/B1 slots to the A2/B2 slots on my system with a 5950. That’s actually the 2nd time I’ve had to do that on Ryzen - my 2600 had the same problem back in the day. Both times the memory was on the QVL.

Tried moving around the dimms, first two dimms and then one, alternating dimms and slots every time. Nothing made any difference and I was never able to boot with xmp speeds enabled. I discovered too that I could very predictably trigger a bluescreen by launching my usual set of VM’s and then opening any h265 media in either one of them.

With only an hour or so to troubleshoot this nonsense on a normal day my patience has now run out and this CPU is taking a trip back to the store.

Also discovered something unexpected in that my 3950x using no thermal paste has lower core-to-core thermal deltas between hottest and coolest cores than what I was seeing with the 5950x (5° vs 8°).

I was a little worried about the temps on my 5950x at first as well, but it’s normal. watching a video for example and doing other random tasks puts mine around 40-48c. If I start up Cyberpunk, it pushes the temp up to 68c for the package temp (the hottest temp measurement for the cpu) Ambient temp in my room is right around 24.4c right now (76f). This is using a IceGiant ProShiphon Elite cooler.

Also mine runs great with DOCP running. Typically if DOCP won’t work it’s a problem between the ram and the motherboard.

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.