DOCP unstable, CMH64GX4M4D3600C18 kit on a ASUS Dark Hero?

[alikabir84]

Having stability issues on a brand new build with CMH64GX4M4D3600C18 Vengence Pro SL ram, system runs fine on default and when I apply the XMP profile the system reboots randomly and getting multiple WHEA errors when there is a BSOD (MINI BUMP GENERATED). Running Cinebench R23 would instantly crash the system when the XMP is enabled the crashes are happening on both single core and multi core. I have since changed to default all auto system settings in bios and ran 6 hours of Prime 95 and 7 hours of Multicore R23 test all have passed. Is this ram kit not compatible with my system?.

Another thing I noticed was the packaging states this RAM kit is compatible for Intel boards where as the description on amazon stated clearly that its optimized for both Intel and AMD?

System Specs:

Asus Dark hero, 5950x, 64gb corsair vengence pro sl, 2xnvmes 980 pro 970 evo plus. 2 x SSD 860-870, 1x HDD WD Purple. Corsair RM1000x power supply, noctua NSD-15S.

 

 

[Corsair Notepad]

Did you make sure to clear your motherboard CMOS before installing the new memory kit? And is your BIOS fully updated? Could be your motherboard is retaining settings from the previous kit which would cause issues such as you are experiencing, or your potentially your BIOS needs updating.

[alikabir84]

Hi, this was a new build so didn't clear the cmos. What I have done now is removed 2 dimms A2 B2 and done the R23 30 minutes test on single core it passed with two dimms and then I ran 2 hour custom R23 single core test its ongoing at the moment and 1 hour has passed. when this stability test completes I will test the remaining dimms on the same slots.

The test are done on all auto settings in bios which is updated to 4201 Asus Darkhero non beta. I performed a 7 hours R23 MULTI CORE test which PASSED (all default settings ram running at 1064Mhz) on all 4 DIMMS  64gb installed. Should I clear cmos now would that help?

 

[alikabir84]

System randomly restarted on multiple occasions its been three days since I built this system errors started popping up with in 4 hours of windows installation and benching. 7 hours of Prime95 also all 64GB at default settings passed. Its when on DOCP profile that this system started with BSODs and random reboots no overclock what so ever just XMP profile enabled. 

[Corsair Notepad]

What happens if you enable DOCP but then manually change the speed from 3600mhz to the officially supported DRAM speed of that CPU at 3200mhz? Could be an issue with the Integrated Memory Controller on the CPU not able to run those modules at 3600mhz with all 4 DIMMs populated as that puts a lot of stress on the IMC when all 4 DIMMs are populated.

[alikabir84]

I haven't tried that , I have passed 3 hours R23 stability test on two Dimms A1, B1 populated going to try the other two on the same slots to see if its the ram dimms then will try 3200Mhz as you suggested.

[alikabir84]

Also like to point out I had the settings at default running at 2133 (1064Mhz) with all 4 dimms slots populated. Passed 7 hours for multi core then failed within an hour of Single core R23 (2133Mhz default) I am now testing the remaining two dimms in the same slots. Whats the reason to check them at 3200Mhz when they don't even pass at 2133?

[Corsair Notepad]

Cinebench R23 is a CPU test and is not indicative of memory stability or functionality in the end. As your initial complaint was with DOCP enabled and you only mentioning that you had them set to 2133mhz one time before and them passing my reply was, and still is a valid point. Also if you are only testing two modules you should be inserting those modules in the optimized DIMM slots for a 2 stick configuration. Based on your statements when you were doing so you had them installed in the 1st and 3rd slot (A1/B1), which according to your motherboard's manual you should have them in the 2nd and 4th (A2/B2).

Installing memory in the incorrect DIMM slots as you stated you did can also lead to stability issues.

[alikabir84]

I understand and I did just that , just finished a full 2 hour test on the A1 B1 SLOTS no issue the R23 test passed then I removed the same working sticks and shifted to B2 A2 and did only a 30 minute test and within 10 minutes the system crashed with another WHEA error 32 GB installed. What seems to be the problem now as I was running in the correct configuration.

[Corsair Notepad]

If they are the same modules and all you did was swap the DIMM slots then you have bad DIMM slots.

[alikabir84]

Yep that is what I was leaning towards will know for sure as I am now on one dimm and on B2 no crashed yet....could this be a pressure mounting issue I have a noctua NSD-15S? If it was a bad dimm slot it shouldn't detect or run the ram in the first place ?

Edited June 2, 2022 by alikabir84

[alikabir84]

yep instant crash on A2 bad slot then?

 

[alikabir84]

Testing with trident Z from my 3900X rig to confirm.

[alikabir84]

Corsair ram is good bad slot as it just failed on Trident Z as well. Thanks for the help.

[Corsair Notepad]

Glad to hear that you were able to confirm the issue, but of course sorry to hear at the same time since replacing a motherboard in a built system isn't the most fun thing to do.

[alikabir84]

RAM slots are all good the Processor is the problem, swapped it with a 3900x and all slots are working fine no crash or hang , have to RMA the Processor. was not expecting these Corsair vengeance Pro sl to run at 3600Mhz on a 3900x  really surprised its a picky processor.

[RKaySwiss]

I had such a WHEA EventLog Spammer last year too, WHEA Errors from the first minute, but no crashes, just hundred thousands of L2 Cache Errors (but all corrected) and a few L1 Cache Errors (BSODs), the **** thing about it in my case, as the crashes very rare, i had to wait for the sendin to intel and their checks, before the vendor replaced it - in your case, the problem can be quickly be demonstrated and is clear, so replacement shouldn't take that much time!

[alikabir84]

Yes I eventually sent it in for a replacement, in the mean time I borrowed a 5950X from a friend to check  the motherboard among other things and this kit is just not stable at all, why would corsair market this as compatible with AMD. This is a new B2 stepping of 5950x, would not work with this kit at 3600Mhz? it initially booted fine and then I noticed weird clitches in game, CRC errors on software ,  corruption on data sizes . Ran memtest and within 30 minutes got errors on first pass( CLEARED CMOS) , ran 3200Mhz all pass 9 hours , then tried 3533Mhz 9 hours pass and then fail within a couple of days with random reboots on idle and a new error where the board reboots to a Q code 00 with Dram led lit up orange, I have to manually switch off the power supply and then it boots fine?.

So far running this kit at default has worked fine 2133Mhz , run games fine, multiple render sessions of 9+ hours Blender all fine.  At this point I am certain its just a memory compatibility issue dual rank must take a hefty toll on the IMC or the kit is just not compatible with Asus Dark Hero. 

Edited July 16, 2022 by alikabir84