Can't enable XMP profile for QVL listed RAM (Asus Z370-E)

[CalvinHobbesN7]

Greetings, kind people of the internet!

 

Everywhere I go, everything I watch, it seems that enabling XMP is a fast, stable, and easy way to get RAM running at its rated speed. Unfortunately enabling the default XMP profile causes the system to become unstable.

 

Things to get out of the way:

• I carefully chose the RAM for my build from the official motherboard's QVL.
• The UEFI BIOS has been updated to the newest release, currently 0616.
• I can manually OC the ram to 16-18-38, 1.353V, and 2533MHz for a stable boot - but this is the maximum I've been able to achieve for a stable boot.
• I have manually attempted a 17-19-39, 1.36V, and 3200MHz boot, but this was also unsuccessful.
• In order to restart the system after raising the bar, I have to flip the PSU physically off, wait 10-15 sec, then boot - pressing F2/Del will open up the dialogue stating that a previous bios change may have caused instability.
• I have successfully run memtest86 at 16-18-38, 1.353V, 2533MHz, without any errors.
• SEPARATELY: I have successfully overclocked this processor to 5GHz and it is quite stable. Before attempting XMP or any ram changes, the BIOS is reset to default and there is no CPU overclock applied during my tests.

Relevant system hardware

• MB: Asus STRIX Z370-E
• CPU: Intel i7-8700K
• RAM: F4-3200C16Q-32GTZSW

Any information, tips, or thoughts on what the issue might be are most welcome.

 

Thank you!

 

-CalvinHobbesN7

[KeitaRR]

maybe RMA your ram and try a new set?

[racer243l]

2533? Thats even below the 8700k´s standard 2666. Enable XMP and drop it to say 3000Mhz and see if it boots

[CalvinHobbesN7]

(Altered post below)

Edited May 31, 2018 by CalvinHobbesN7

[CalvinHobbesN7]

36 minutes ago, KeitaRR said:

maybe RMA your ram and try a new set?

I would RMA this RAM except that it passed the Memtest86 without a single error.

33 minutes ago, racer243l said:

2533? Thats even below the 8700k´s standard 2666. Enable XMP and drop it to say 3000Mhz and see if it boots

 

That's... really surprising to hear and I wasn't aware of this. The default BIOS settings are even lower at 2133 or something. I can't get anything to work higher than 2533. Period. I took it up one notch at a time until it failed - 2600 is the first failure.

[racer243l]

12 minutes ago, CalvinHobbesN7 said:

I would RMA this RAM except that it passed the Memtest86 without a single error.

 

That's... really surprising to hear and I wasn't aware of this. The default BIOS settings are even lower at 2133 or something. I can't get anything to work higher than 2533. Period. I took it up one notch at a time until it failed - 2600 is the first failure.

2133 is DDR4´s default speed, the 8700k´s memory controller is rated for 2666. And that really is supprising. Your memory is rated to run at 3200 speed, so passing memtest at 2666 isn´t ralevent. RMA it.

[CalvinHobbesN7]

1 minute ago, racer243l said:

2133 is DDR4´s default speed, the 8700k´s memory controller is rated for 2666. And that really is supprising. Your memory is rated to run at 3200 speed, so passing memtest at 2666 isn´t ralevent. RMA it.

Do we know it's the RAM? Or could it be the processor (I've been successful at 5GHz OC)? Or could it be the motherboard? How do I know where the problem lies?

[racer243l]

1 minute ago, CalvinHobbesN7 said:

Do we know it's the RAM? Or could it be the processor (I've been successful at 5GHz OC)? Or could it be the motherboard? How do I know where the problem lies?

Well, it could also be the CPU´s memory controller, some wierd board or BIOS issue. It´s a long list. You can try each stick of the kit and all slots on your motherboard individually and see if it changes anything

[i_build_nanosuits]

1 hour ago, CalvinHobbesN7 said:

Greetings, kind people of the internet!

 

Everywhere I go, everything I watch, it seems that enabling XMP is a fast, stable, and easy way to get RAM running at its rated speed. Unfortunately enabling the default XMP profile causes the system to become unstable.

 

Thank you!

 

-CalvinHobbesN7

XMP profile is considered as overclocking and depending on the manufacturer's validation testing it's not impossible to encounter some RAM kits that can't work at the advertised XMP clocks...especially if you are using more than a 2 stick kit...that's not uncomon to see this issue...you got unlucky...

 

14 minutes ago, CalvinHobbesN7 said:

Do we know it's the RAM? Or could it be the processor (I've been successful at 5GHz OC)? Or could it be the motherboard? How do I know where the problem lies?

Problem could be due to the IMC on your CPU being sub par...or, more likely...Gskill quality control being a bit too loose in shipping memory that just barely made the cut for the speed and you using 2 kits of these togheter creates instability...motherboard has nothing to do with it IMO...you could always update your board bios to the latest version to clear this one out of the way...but my guess really is down to memory quality control...what you have there is probably a couple 3000mhz sticks trying to pass for 3200mhz which they probably passed a couple validations on their own but once combined with other sticks and used in a less than ideal situation might see instability.

 

Hope this help!

[CalvinHobbesN7]

47 minutes ago, i_build_nanosuits said:

...you using 2 kits of these togheter creates instability...motherboard has nothing to do with it IMO...you could always update your board bios to the latest version to clear this one out of the way...but my guess really is down to memory quality control...what you have there is probably a couple 3000mhz sticks trying to pass for 3200mhz which they probably passed a couple validations on their own but once combined with other sticks and used in a less than ideal situation might see instability.

I think that's interesting - I'm only using a single kit that contains 4x ram sticks so I'm not personally combining any kits together. Just using what G.Skill sold me. I'll get in touch with G.Skill to begin the RMA process, sounds like a pain in the ass.

[CalvinHobbesN7]

56 minutes ago, racer243l said:

Well, it could also be the CPU´s memory controller, some wierd board or BIOS issue. It´s a long list. You can try each stick of the kit and all slots on your motherboard individually and see if it changes anything

What should I look for? See if individual sticks can make the 3200 speed? Also - you recommend trying each slot, and not just the MB recommended slot for running a single stick? That's 16 tests...

[badreg]

Since you had issues with your previous RAM as well, the culprit is starting to look like either a motherboard issue or a faulty memory controller in your CPU.

[racer243l]

14 hours ago, CalvinHobbesN7 said:

What should I look for? See if individual sticks can make the 3200 speed? Also - you recommend trying each slot, and not just the MB recommended slot for running a single stick? That's 16 tests...

I know a lot of work. Try each stick in the recommended slot first and see if one get´s you to 3200Mhz.

10 hours ago, badreg said:

Since you had issues with your previous RAM as well, the culprit is starting to look like either a motherboard issue or a faulty memory controller in your CPU.

I point in this direction as well...

[CalvinHobbesN7]

7 hours ago, racer243l said:

I point in this direction as well...

Fun story - I actually got both XMP and manual overclocking to function. The weird thing is I had to remove my GPU to do it. I created a new forum post regarding this new-found knowledge. Short version: I think it may be the riser cable I'm using, and others feel it may be a voltage problem. I think the two solutions (increasing voltage + better riser cable) are connected.

 

Thank you for your input here!

[badreg]

4 hours ago, CalvinHobbesN7 said:

Fun story - I actually got both XMP and manual overclocking to function. The weird thing is I had to remove my GPU to do it. I created a new forum post regarding this new-found knowledge. Short version: I think it may be the riser cable I'm using, and others feel it may be a voltage problem. I think the two solutions (increasing voltage + better riser cable) are connected.

 

Thank you for your input here!

Why are you using a riser cable in the first place? In almost all cases, this leads to a significantly worse thermal outcome than slotting the card directly into the motherboard. 

[CalvinHobbesN7]

8 minutes ago, badreg said:

Why are you using a riser cable in the first place? In almost all cases, this leads to a significantly worse thermal outcome than slotting the card directly into the motherboard. 

Aesthetics, since I'm using an RGB EKWB block to cool things down, and since Linus didn't have a problem using them to extend out to over 3.0 meters either (granted I'm using a far less superior one, I've since ordered a premium one used on his video at 30cm).

[badreg]

Ah, got it. I didn't remember you saying that you were water cooling. I hope the new riser cable works out.

[racer243l]

17 hours ago, CalvinHobbesN7 said:

Fun story - I actually got both XMP and manual overclocking to function. The weird thing is I had to remove my GPU to do it. I created a new forum post regarding this new-found knowledge. Short version: I think it may be the riser cable I'm using, and others feel it may be a voltage problem. I think the two solutions (increasing voltage + better riser cable) are connected.

 

Thank you for your input here!

Good to hear you found a solution. Crazy how PC tech can be influenced by each other

[CalvinHobbesN7]

On 6/2/2018 at 1:51 AM, racer243l said:

Good to hear you found a solution. Crazy how PC tech can be influenced by each other

Even the smallest of things like a poor riser-cable can be affected by ram frequency. it's certainly been a learning experience!