• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Best RAM settings

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

kelsinni

Registered
Joined
Jan 13, 2013
Hello all,
After recently updated my system from my old Phenom II X6 1075T to my new AMD Ryzen 5 2600, I'm trying to find a decent setting to my new DDR4 RAM...

I have 2 different KIT of 16Gb each so I know that I will have more difficult to get both stable with an high frequencyÂ…

First I tried just the 2x8Gb G.Skill Trident Z RGB (F4-3200C16D-16GTZR) @ 3200 Mhz CL 16-18-18-38 2T @ 1,35v with the XMP profile activated but I couldnÂ’t pass more than a few minutes in Memtest.

Then I tried just the 2x8Gb TEAMGROUP T-FORCE Gaming DELTA RGB (TF4D416G3200HC16CDC01) and with the XMP profile I got a stable 3200Mhz CL 16-18-18-38 2T @ 1,35v.

Now I have both working and till now the best result I have managed was @ 2933 CL 14 17 17 38 2T @1,35v.
Screenshot-1.png

IÂ’m here trying to know about your expertise and opinion about if this last value is already a good mark regarding my hardware or, if you think that I have more room to go on and push a little bit and get them furtherÂ…

Thanks in advance for your help.

Cheers!
 
Try a little vSOC increase. I used 1.1v for my 2600x. https://www.overclockers.com/forums/showthread.php/781682-AMD-Ryzen-Tool-Kit. If not, really 2933 should give you nearly the same performance, especially with those timings.

Edit are you using two different kits together in 4x8GB? If so it might be beneficial to download taiphoon burner which will let you see what specific IC the kits are using. They may be the same or may be different. It might not be super relevant since your not going high with the memory clock, but it would still be interesting to know.
 
Try a little vSOC increase. I used 1.1v for my 2600x. https://www.overclockers.com/forums/showthread.php/781682-AMD-Ryzen-Tool-Kit. If not, really 2933 should give you nearly the same performance, especially with those timings.

Edit are you using two different kits together in 4x8GB? If so it might be beneficial to download taiphoon burner which will let you see what specific IC the kits are using. They may be the same or may be different. It might not be super relevant since your not going high with the memory clock, but it would still be interesting to know.

Thanks for your reply...
Yes, I'm using 2x8Gb+2x8Gb (32Gb) and in fact they are different...one is Hynix and other Samsung
Screenshot-1.png Screenshot-2.png
My goal isnt pushing them to their maximum possible values... just to get an ideia if this value is already a good one or if I have plenty more space to move on...

Because to be honest, Im still attached to the DDR3 age and so this is all new for me...

Thanks again

Cheers!
 
To be honest, my initial PC setup config was to have just 16Gb RAM and so I got 1st the G.Skill Kit but then I got a very very good deal and got that opportunity to increase to 32Gb...

I had precisely that idea about different RAM kits but... I tought that if I could get a stable 32Gb even @ 2933 CL14 was maybe better than 16Gb @ 3200 CL16.

So, in your opinion it will be advisable to eventually sell one of the Kits and try to get another equal? If so which of the two you think its better? G.Skill ou Teamgroup?

Thanks
 
Matching sets are always better.
The Trident Z's are tried and true. You can get very good results with them. You just have to tweak a little instead of going straight XMP.
 
You're probably right...

But a strange thing that happend was that both RAM have exactly the same specs regarding frequency, latency and voltage... and I got 3200 CL16 at the 1st try with the Teamgroup kit and never acheived that same frequency with the G.Skill... It is strange or not exactly?

In that case, your advice is to sell the Teamgroup kit and try to get an equal G.skill kit right?

Thanks again...
 
Well here's another question for you, are you actually utilizing 32GB or did you just want to have it? Consider just running one kit if you don't actually need 32gb. Unused memory does not make your system faster. Maybe the trident z could do better with manual tweaking, you could always try both by themselves and see which does you better, but if the teamgroup worked with XMP and that's what you want, consider keeping it. But also consider that buying the same spec kit, even within the same brand, is not guaranteed to get you matching kits. Especially not with something as common as 3200 CL 16.

I'm not sure if you're saying that adding the trident wouldn't work xpm with all four sticks or if you already tried them independently. Keep in mind 4x8GB is a lot more work for the memory controller than 2x8GB, and 4 vs 2 stick performance is also impacted by motherboard design.
 
Well here's another question for you, are you actually utilizing 32GB or did you just want to have it? Consider just running one kit if you don't actually need 32gb
In fact I think that most of the time im not using the 32Gb... I mostly use Photoshop and regarding games, I just play them on Bluestacks (PC android emulator).
I just get the 2nd 16gb kit because it was very similar to the 1st (Frequency, timings, voltage,...) and I tought that would be the most important...

Unused memory does not make your system faster. Maybe the trident z could do better with manual tweaking, you could always try both by themselves and see which does you better, but if the teamgroup worked with XMP and that's what you want, consider keeping it. But also consider that buying the same spec kit, even within the same brand, is not guaranteed to get you matching kits. Especially not with something as common as 3200 CL 16.
Ok got it... even I get an equal RAM kit I can have the same issues...

I'm not sure if you're saying that adding the trident wouldn't work xpm with all four sticks or if you already tried them independently. Keep in mind 4x8GB is a lot more work for the memory controller than 2x8GB, and 4 vs 2 stick performance is also impacted by motherboard design.
With the Trident I couldnt never acheived the the 3200Mhz... 1st I tried using just the XMP and then tweaked a little bit with a little increse in the SoC voltage but even so... didnt work. Seems that these G.Skill kit doesnt like my CPU-Mobo combination or doesnt like to work at 3200 :)

Thks
 
Is that with the trident by itself or both kits? It could be that, and it could be that as mr scott said the ram will do it, there is just a timing that isn't jiving. I think your best option is to either keep what you have or go back to the 16gb setup if you're not using the 32gb capacity. You can always keep the kit and know for a mild performance cost you can go back to 32gb if it was really a good deal.
 
Is that with the trident by itself or both kits? It could be that, and it could be that as mr scott said the ram will do it, there is just a timing that isn't jiving. I think your best option is to either keep what you have or go back to the 16gb setup if you're not using the 32gb capacity. You can always keep the kit and know for a mild performance cost you can go back to 32gb if it was really a good deal.

The 3200Mhz I just got it with the Teamgroup kit alone... With the G.Skill alone and with both, I couldnt do it.

Seems really like that the G.Skill doesnt like one of the timings but, its quite strange too because all the timings used are exactly the ones that are mentioned on thae respective XMP of them...

For now I got 2933Mhz CL 13 17 17 38 @1,35v with both kit (32Gb) installed.

Im now gonna try lower the 38 to 36 and see if it continues ok...
 
Nice! I think the most important thing is to have fun. Honestly that's pretty impressive all things considered. When it comes to memory (and I really am not the expert), there are timings upon timings. There are primary timings, secondary timings, timings that the BIOS doesn't even expose to the user that get set when the computer first boots or has a BIOS reset (correct me if I'm wrong to call what is set during memory training timings). In other words there are a lot of variables so getting two different kits, even with apparently similar timings, to play as well as you have is quite impressive.
 
Back