Is this a normal RAM behavior?

[Jakub_NF]

I just swapped out the motherboard and CPU, the PC booted fine so I went to apply some settings in BIOS like fans/XMP. The PC booted fine with ram at its rated speed (Corsair Vengeance 2x8GB 3200MHz CL16). I always had the ram running at 3600MHz on my old system (Gigabyte Z390 gaming x, i5-8600k) without changing any other memory settings. When I tried the same speed on the new system (Aorus Elite Z690 AX, i5-12600KF) the mobo just showed DRAM error and wouldn't boot. After waiting for 10 mins I restarted it but the bootdrive got bricked and I couldn't fix it even with the windows media creation tool so I had to reinstall. Is it possible that the ram will just work differenty on this system and won't overclock? 

[LIGISTX]

20 minutes ago, Jakub_NF said:

I just swapped out the motherboard and CPU, the PC booted fine so I went to apply some settings in BIOS like fans/XMP. The PC booted fine with ram at its rated speed (Corsair Vengeance 2x8GB 3200MHz CL16). I always had the ram running at 3600MHz on my old system (Gigabyte Z390 gaming x, i5-8600k) without changing any other memory settings. When I tried the same speed on the new system (Aorus Elite Z690 AX, i5-12600KF) the mobo just showed DRAM error and wouldn't boot. After waiting for 10 mins I restarted it but the bootdrive got bricked and I couldn't fix it even with the windows media creation tool so I had to reinstall. Is it possible that the ram will just work differenty on this system and won't overclock? 

Yes, that is entirely possible. It’s also possible the RAM isn’t and wasn’t actually stable at those speeds without tweaking voltages or latencies. 

 

RAM OCing is the most difficult form of OCing, it’s hard to get stable, and it’s even more difficult to test… 

 

I am not trying to discourage you from trying it, but make sure you properly and fully test for stability; normal CPU stability tests are not at all good enough. 
 

Fully read and understand this GitHub page, and use the tools they recommend the validate. 
 

https://github.com/integralfx/MemTestHelper/blob/oc-guide/DDR4 OC Guide.md

 

But, the short version is yes, RAM can behave differently in different platforms. My RAM was rock solid at 3600 on my 9900k system, I had to drop it down to 3500 on my 10700k system, all other settings being identical. It seemed stable under all standard tests, but after 8+ hours of Google’s RAM burn in test which runs on linux, I started to see an error or two crop up. 

[Pandur]

Any overclock is a gamble. You can have two identical systems (not even different generations like you do) and the same stick of memory will work fine on one, but not the other. No two IMC have the exact same characteristics. This is exactly the same as overclocking CPU frequency.

 

It is possible that you just need a bit more voltage to the IMC, which would be System Agent voltage in most bios I have worked with. But it could also mean that your old overclocked settings are just not correct for the new platform.

 

LIGISTX have some homework for you in the post above. 

[Jakub_NF]

13 hours ago, LIGISTX said:

Yes, that is entirely possible. It’s also possible the RAM isn’t and wasn’t actually stable at those speeds without tweaking voltages or latencies. 

 

RAM OCing is the most difficult form of OCing, it’s hard to get stable, and it’s even more difficult to test… 

 

I am not trying to discourage you from trying it, but make sure you properly and fully test for stability; normal CPU stability tests are not at all good enough. 
 

Fully read and understand this GitHub page, and use the tools they recommend the validate. 
 

https://github.com/integralfx/MemTestHelper/blob/oc-guide/DDR4 OC Guide.md

 

But, the short version is yes, RAM can behave differently in different platforms. My RAM was rock solid at 3600 on my 9900k system, I had to drop it down to 3500 on my 10700k system, all other settings being identical. It seemed stable under all standard tests, but after 8+ hours of Google’s RAM burn in test which runs on linux, I started to see an error or two crop up. 

I think it's the 8-9th gen platform that is just really good for that. My friend had the same cpu with a z370 asus mobo and was able to just up the ram by 400MHz without changing anything. I ran that setup for over 2 years and never had any issues so I'm assuming it was stable.

[LIGISTX]

6 hours ago, Jakub_NF said:

I think it's the 8-9th gen platform that is just really good for that. My friend had the same cpu with a z370 asus mobo and was able to just up the ram by 400MHz without changing anything. I ran that setup for over 2 years and never had any issues so I'm assuming it was stable.

If you didn’t check stability, don’t assume it’s stable. RAM OCing isn’t for the faint of heart and it isn’t a good idea unless you take the time to understand and learn what your doing. 
 

RAM OCing is one of the far overclocking types that can actually cause issues, like slowly corrupting your data, cause crashes many months or years down the line that don’t make any sense and don’t seem like they would be RAM related, just generally can cause issues you don’t want to deal with. 
 

If your going to OC RAM, spend the time to learn it and do it correctly, and most importantly, correctly and thoroughly validate for stability. At the end of the day your chasing 2-3% performance gains at best, it’s not worth hosing up your system over assuming it’s stable just to get an inconceivable performance bump. 

[Jakub_NF]

On 5/20/2022 at 3:49 PM, LIGISTX said:

If you didn’t check stability, don’t assume it’s stable. RAM OCing isn’t for the faint of heart and it isn’t a good idea unless you take the time to understand and learn what your doing. 
 

RAM OCing is one of the far overclocking types that can actually cause issues, like slowly corrupting your data, cause crashes many months or years down the line that don’t make any sense and don’t seem like they would be RAM related, just generally can cause issues you don’t want to deal with. 
 

If your going to OC RAM, spend the time to learn it and do it correctly, and most importantly, correctly and thoroughly validate for stability. At the end of the day your chasing 2-3% performance gains at best, it’s not worth hosing up your system over assuming it’s stable just to get an inconceivable performance bump. 

I ran a stability test for 24 hours with all the overclocking I did after setting up the PC and it was fine. I definitely don't know what I'm doing but if I get lucky and it works then I'm happy.

[LIGISTX]

3 hours ago, Jakub_NF said:

I ran a stability test for 24 hours with all the overclocking I did after setting up the PC and it was fine. I definitely don't know what I'm doing but if I get lucky and it works then I'm happy.

What did you use to test for RAM stability? Most tests folks recommend for CPU will not touch RAM in any meaningfulI way.

 

I promise, not correctly checking for RAM stability is not worth the few % gains. RAM OCing is the only real OC that shouldn't be YOLOed. CPU OCing is relatively easy and safe, GPU OCing is literally a no brainer and is almost 100% safe regardless of what you do.

[Jakub_NF]

10 hours ago, LIGISTX said:

What did you use to test for RAM stability? Most tests folks recommend for CPU will not touch RAM in any meaningfulI way.

 

I promise, not correctly checking for RAM stability is not worth the few % gains. RAM OCing is the only real OC that shouldn't be YOLOed. CPU OCing is relatively easy and safe, GPU OCing is literally a no brainer and is almost 100% safe regardless of what you do.

I used aida64 for the long test and later did 6 or so hours of prime95 

[LIGISTX]

2 hours ago, Jakub_NF said:

I used aida64 for the long test and later did 6 or so hours of prime95 

I would visit the ddr4 overclocking GitHub page and use some of the stress tests they mention. If it passes, your good to go.