z690 Apex with Samsung DDR5 Memory Issues?

[Murlo26]

Hi,

I have a new build put together and the memory is giving me all sorts of issues. I have seen a few people saying that the Apex board might have some dim slot issues but not sure if that is actually an issue. I snagged up a 6000mhz CL36 G Skill kit and tried pairing with my apex board and 12900k. Everything is on a custom water loop with fairly solid temps. I am on the latest bios. I recognize the Samsung flavor of sticks are not officially supported by the Apex board yet, but am curious as to why (per the QVL). The hero board has mine listed already but the apex list seems light. I tried the XMP profile and have had very little luck. I have been able to get it to boot into windows but typically crashes on any heavy workload and sometimes even just being in windows things lock up.

I am just curious if this is something I need a bios update for and just have to wait or I should look at going after different memory. Part of why I got the apex was to finally play around with memory overclocking but not even being able to run xmp kind of sucks. Just surprised that the apex board has so little support for memory so far.

Any thoughts?

I appreciate any help and sorry if this is covered elsewhere.

PC Components:

12900k (stock for now until I get memory sorted)
z690 Apex
G Skill F5-6000U3636E16GX2-TZ5K (32gb)
EVGA 3090 Kingpin Hydrocopper
EVGA 1300W G2 PSU
Corsair 5000d
2x360mm radiators
Windows 11

Edit: I have also tried the built in profiles on the apex, like all of them and none of those seem to work even speeds down to like 5200mhz, wouldn't boot into windows on those. I also tried the XMP and bumped voltage up...no luck really.

[Silent_Scone]

Murlo26 wrote:
I personally have never had issues with XMP profiles being 100% stable using asus products for like a decade. I completely understand what you are saying though. I was just surprised that there is really no samsung dims listed on the QVL for apex but it's all there for the hero board.

I appreciate that link, I will read though it and probably work on getting my sticks to work.

I guess it just seems like after some research that the SK hynix stuff has been more stable for people so I was tempted to get some different sk hynix stuff to make life easier.

Thanks again, I will read though it.

PS. I am still interested to hear why the apex QVL is fairly bare and way lower than a board like the APEX. I am guessing it's just because the hero is a much more mainstream board that probably sells FAR more than the apex and people that get the apex are expected to have to tweak their own stuff. But again, I have never really had issues in the past.

Generally speaking 2 DIMM boards are less popular, so memory vendors are less inclined to validate as many kits as possible. That said, as many are now aware the memory topology of the Apex makes overclocking considerably easier.

This does not mean it’s exempt from certain combinations of CPU samples and kits not jiving well with auto rules, or weak memory controllers. XMP hasn't ever guaranteed stability and it certainly doesn't on Z690. Majority of CPU should be able to do 6000 2T, however, adjustments may need to be made. Having never personally experienced an issue before doesn’t mean one doesn’t exist, right? This is no different, and comes down to signal alignment. For most of us, it’s better we accept these things for what they are and tune accordingly.

Some of these voltages are sensitive to change. Each CPU is different, so guides will only get you so far. Sometimes all one can say is it works on my machine if I use X&Y. This isn’t anything new, either. These things have always been there at varying degrees depending on the platform and how ambitious one is when aiming for frequency. On Z690, the window for stability with some of these signal voltages is quite narrow.


To start:

*Bump VDD and VDDQ to 1.35v

*VCCSA 0.98c

MCVDD 1.25 (Extreme Tweaker>Advanced Memory Settings)

[Murlo26]

Silent Scone@ROG wrote:
Generally speaking 2 DIMM boards are less popular, so memory vendors are less inclined to validate as many kits as possible. That said, as many are now aware the memory topology of the Apex makes overclocking considerably easier.

This does not mean it’s exempt from certain combinations of CPU samples and kits not jiving well with auto rules, or weak memory controllers. XMP hasn't ever guaranteed stability and it certainly doesn't on Z690. Majority of CPU should be able to do 6000 2T, however, adjustments may need to be made. Having never personally experienced an issue before doesn’t mean one doesn’t exist, right? This is no different, and comes down to signal alignment. For most of us, it’s better we accept these things for what they are and tune accordingly.

Some of these voltages are sensitive to change. Each CPU is different, so guides will only get you so far. Sometimes all one can say is it works on my machine if I use X&Y. This isn’t anything new, either. These things have always been there at varying degrees depending on the platform and how ambitious one is when aiming for frequency. On Z690, the window for stability with some of these signal voltages is quite narrow.


To start:

*Bump VDD and VDDQ to 1.35v

*VCCSA 0.98c

MCVDD 1.25 (Extreme Tweaker>Advanced Memory Settings)

So I can find VDD and VDDQ easy enough, I am not seeing the other two. What units are the VCCSA in, the "c"?

Am I missing something? Is MCVDD the memory controller voltage?

[Murlo26]

Murlo26 wrote:
So I can find VDD and VDDQ easy enough, I am not seeing the other two. What units are the VCCSA in, the "c"?

Am I missing something? Is MCVDD the memory controller voltage?

Still was hoping to get an answer on this.Â*

[Silent_Scone]

Murlo26 wrote:
Still was hoping to get an answer on this.Â*

MCVDD is the Memory Controller Voltage.

VCCSA is CPU System Agent Voltage.

[Murlo26]

Silent Scone@ROG wrote:
MCVDD is the Memory Controller Voltage.

VCCSA is CPU System Agent Voltage.

Awesome, thank you! I will try this later.

[lLegendaary]

I might upgrade my setup to Z790 and skip Z690
----~> Skipping Z690 [ because a lot of complaina nd complete new generation most vendors facing issue in stability and support stuff you will see much more stability OC' starting with Z790 ending to Z890


Scenario here Z890 ----~> will have much more ability to OC With Advanced GDDR5x Memory and Much more support for Fully Support X16-PCIE5.0 256Gbps-Support
Z790 ----~> will have accepted OC and more stability than Z690

Please be patient don't rush in Things Z690 is just introduce new technologies and more unlocked hardware stuff like RAMS DDR5 advanced and PCIE5.o and bandwidth plus lane

[MoKiChU]

Hi,

XMP I and XMP II do not correspond to XMP version, but to XMP mode :

XMP I mode = ASUS optimized XMP profil
XMP II mode = Memory manufacturer original XMP profil

XMP version depend only on the generation of memory used :

[Phroste]

MoKiChU wrote:
Hi,

XMP I and XMP II do not correspond to XMP version, but to XMP mode :

XMP I mode = ASUS optimized XMP profil
XMP II mode = Memory manufacturer original XMP profil

XMP version depend only on the generation of memory used :

That does not explain why the XMP 3 3rd profile along with the 2 custom/renamable profiles are missing in the bios. Since the 3rd XMP 3 profile and the 2 custom profiles are missing, it does not seem like the Bios recogizes it yet, since they are already encoded on the memery dimms, which is why I was thinking the Asus bios is fully optimized for DDR5 XMP 3.0 just yet. My thoughts anyway.

[MoKiChU]

Phroste wrote:
That does not explain why the XMP 3 3rd profile along with the 2 custom/renamable profiles are missing in the bios. Since the 3rd XMP 3 profile and the 2 custom profiles are missing, it does not seem like the Bios recogizes it yet, since they are already encoded on the memery dimms, which is why I was thinking the Asus bios is fully optimized for DDR5 XMP 3.0 just yet. My thoughts anyway.

Hi,

ASUS BIOS/UEFI has nothing to do with this, the Intel slide that I quoted to you indicates what XMP 3.0 allows, it is then up to the memory manufacturer to do what is usually done with XMP 1.0/2.0 historically (1 or 2 static profiles and nothing else), or to use the possibilities offered by XMP 3.0 (and if so, these will be visible in your ASUS BIOS/UEFI).

[Murlo26]

MoKiChU wrote:
Hi,

ASUS BIOS/UEFI has nothing to do with this, the Intel slide that I quoted to you indicates what XMP 3.0 allows, it is then up to the memory manufacturer to do what is usually done with XMP 1.0/2.0 historically (1 or 2 static profiles and nothing else), or to use the possibilities offered by XMP 3.0 (and if so, these will be visible in your ASUS BIOS/UEFI).

I am following what you are saying but I am not sure ASUS's BIOS has this option currently, but I haven't really looked. When I use CPUz I can see the different profiles for my memory, I think 3 of them. But the only XMP setting is the max setting one I believe. I could be wrong though as I haven't looked very hard yet. It's still a brand new deal so might not be implemented yet.