Manually Overclocking Corsair Dominator DDR5?

BigSykeDaddy

BigSykeDaddy

BigSykeDaddy

Associate
Joined
6 Dec 2021
Posts
8
Location
Florida
Hello everyone!

First of all I wanna say you guys are awesome! I've enjoyed reading through these forums off and on over the years and I have learned a lot!

I believe I only have a couple posts on these forums so I am still a complete newbie when it comes to overclocking.

I recently "finished" my new build and I started out with one 32gb (2x16) 5200 c40 Dominator Platinum paired with an i9 12900k on a Z690 Hero Motherboard. I mainly use this computer with Adobe programs (Premiere / Photoshop / Audition / After Effects / Illustrator etc) so I decided to grab another 32gb DDR5 kit (same 5200 c40 Dominator Platinum kit) to run 64GB of DDR5.

I have been into tech since messing around with a computer running DOS so I know my way around a little bit, but I am far from an expert. After all these years I NEVER knew that if you purchased the same kit, but in different boxes that XMP might have stability issues (I know, I know, I'll take my noob beating now haha, I have read what you all have to say about people like me = ).

Anyway, when I first got this system up and running with the first 32GB kit I had 0 issues running XMP (obv), and I would like to manually OC these 2 kits since XMP is not stable for me. I have been watching videos and reading some manual OC guides, but I am aware that DDR5 is a little difference since the PMIC chip is on the DIMM itself as opposed to on the mobo like it used to be.

I am not looking for any crazy results, just something a little better than 4000mhz. I tried following one guide and my computer immediately BSOD'd so I am coming to you OC pros for a little guidance. I have learned a little along the way in terms of manually OCing DDR5, so I understand the concept and SHOULD be able to take any advice I receive and hit the ground running. I know that it can be a fine art balancing the voltages and timings etc, but I am willing to learn.

TLDR: <-- OC noob who after 30 years of tinkering with tech did not know trying to run 2 kits (2x16 5200 c40 DDR5 x 2) in XMP would cause instability. Willing to learn, but I really don't want to mess anything up even though modern day computers have a lot of safety features. Already tried XMP 1 and 2 and no dice.

- Thank you all very much in advance and sorry for any typos / run on sentences, I hope everyone is having an awesome weekend!
 
Welcome to the forums, I'm sure someone will have some suggestions for you soon. No need to post the same thing in two sections, I've removed the one from Overclocking.

Good luck on your quest sir :)
 
Tetras said:
If you're still interested you might want to look at the content of buildzoid and der8auer, like this:

A relaxed daily DDR5 OC on the Gigabyte Z690 Aorus Xtreme.
https://www.youtube.com/watch?v=alk3QLyQebs

DDR5 is still very new so I'm not sure you'll get many responses here for now :o
Click to expand...

Hey Tetras, thank you for sending me that link and replying to my thread. I appreciate it! I started watching the video you sent last night and I am looking forward to finishing it this evening.

I tried something new trying to get a my DDR5 to run at speeds greater than 4000mhz. I ended up enabling XMP I in the BIOS and then scrolled down a bit to the DRAM Frequency where I went ahead and dropped the setting it loaded from 5200 to 5000 and left all the other settings the same. So far so good, it passed my "stress" test of running a couple games that used to always crash while trying to use XMP I.

The games did not crash so that's good. From there I rendered a couple demanding videos with a lot of effects on them and that too passed. My next step is to run a proper stress test with OCCT or AIDA64 or something. From the research I did prior to making a post here on these forums it said that some people test for an hour, while others stress test for 24 hours.

Do you have an programs for stress testing that you would recommend over others and in your opinion how long do you think I should let the stress test run before I know it's pretty stable? When I was trying to manually dial in the DDR5 to run at 5200Mhz I was using OCCT to test the OC and the best results I could achieve was OCCT running for the one hour stress test and "only" returning 2 errors. When I looked online to see if it was possible to get by with 2 errors on a stress test and it seems like the consensus is one error is one too many and that those one or two errors could turn into something major down the road to the point where your PC wont start anymore. Again, I am brand new to this and I don't know if that's true or if people are being overly cautious (understandable).

Thanks again for the reply and the link, I hope you have a good day!

-Anthony
 
BigSykeDaddy said:
Hey Tetras, thank you for sending me that link and replying to my thread. I appreciate it! I started watching the video you sent last night and I am looking forward to finishing it this evening.

I tried something new trying to get a my DDR5 to run at speeds greater than 4000mhz. I ended up enabling XMP I in the BIOS and then scrolled down a bit to the DRAM Frequency where I went ahead and dropped the setting it loaded from 5200 to 5000 and left all the other settings the same. So far so good, it passed my "stress" test of running a couple games that used to always crash while trying to use XMP I.

The games did not crash so that's good. From there I rendered a couple demanding videos with a lot of effects on them and that too passed. My next step is to run a proper stress test with OCCT or AIDA64 or something. From the research I did prior to making a post here on these forums it said that some people test for an hour, while others stress test for 24 hours.

Do you have an programs for stress testing that you would recommend over others and in your opinion how long do you think I should let the stress test run before I know it's pretty stable? When I was trying to manually dial in the DDR5 to run at 5200Mhz I was using OCCT to test the OC and the best results I could achieve was OCCT running for the one hour stress test and "only" returning 2 errors. When I looked online to see if it was possible to get by with 2 errors on a stress test and it seems like the consensus is one error is one too many and that those one or two errors could turn into something major down the road to the point where your PC wont start anymore. Again, I am brand new to this and I don't know if that's true or if people are being overly cautious (understandable).

Thanks again for the reply and the link, I hope you have a good day!

-Anthony
Click to expand...

I'd say that it really depends on what you're using the PC for. For example: if you're only gaming and don't have any important data, then "just about stable" is good enough, though do keep in mind that the stability of overclocks can change due to conditions (both workloads and environmental).

In terms of a specific time to run them, I guess that's up to you and how much you rely on the PC, versus the benefit of the extra performance. Usually it's not worth pushing an overclock to the edge, because you get diminishing returns in terms of voltages and stability for the extra Mhz. So, it's normal to find the edge of what you're comfortable with and then dial it down a notch or two, which helps ensure that longer tests won't have any errors.

I would agree though, that any error on a stress test is one too many, because these errors can mean you'll suffer 'silent' corruption to your boot drive (or other data). Though, like I said, if you're just hobbying then maybe you don't care if it will be rock-solid stable for 72 hours.
 
Tetras said:
I'd say that it really depends on what you're using the PC for. For example: if you're only gaming and don't have any important data, then "just about stable" is good enough, though do keep in mind that the stability of overclocks can change due to conditions (both workloads and environmental).

In terms of a specific time to run them, I guess that's up to you and how much you rely on the PC, versus the benefit of the extra performance. Usually it's not worth pushing an overclock to the edge, because you get diminishing returns in terms of voltages and stability for the extra Mhz. So, it's normal to find the edge of what you're comfortable with and then dial it down a notch or two, which helps ensure that longer tests won't have any errors.

I would agree though, that any error on a stress test is one too many, because these errors can mean you'll suffer 'silent' corruption to your boot drive (or other data). Though, like I said, if you're just hobbying then maybe you don't care if it will be rock-solid stable for 72 hours.
Click to expand...

Hey thanks again for the help. I use this PC mostly for work (Adobe Premiere / Photoshop / After Effects / Illustrator / Audition etc etc) and every now and then I'll play some games. I am going with what you and the others said about "one error being one too many" and I am going to play it safe. It has been running stable at 5000 mhz now for almost two days, but I haven't had time to run a stress test (overnight rendering ETC). I think tomorrow before I go to bed I'll let AIDA64 or OCCT run for 8 hours and see what happens. I'm just happy it's running faster than 4000mhz like it was. After all my years messing with tech and building computers this is shockingly the first time I have ran across 2 identical kits purchased a month apart not being able to run XMP without messing with it. I won't forget it though haha so I guess it's a good lesson. As DDR5 matures I may shoot for a 2x32gb 6000+ kit down the line, but I am thrilled moving from 32gb DDR3 to 64gb of DDR5, it's great!

Appreciate the advice and I hope you have a good one!
 
You must log in or register to reply here.
Back
Top Bottom