Okay to mix and match RAM?

Psycho Sonny

Psycho Sonny

Psycho Sonny

Caporegime
Joined
21 Jun 2006
Posts
38,367
I know people will say don't do it but does it actually make a big difference?

I have 2 x 4GB of Crucial Ballistix 2400 and 2 x 8GB of Crucial Ballistix 2400 both DDR4.

Is it okay to use both in their assigned dual channels and let them run at the slower timings of the 2 sets? Obviously the slightly faster set won't be getting used to it's full potential but surely having an additional 8GB of RAM which is 50% more would be worth it?
 
Can't really comment without knowing what platform you're talking about, but there's reasons why it potentially both will and won't work for you.

If running them at default settings, you shouldn't have an issue.
 
hornetstinger said:
no problem I've got a mix of ocz and corsair, 2 and 3GB modules, different speeds, latency etc. I've just set it to lowest speed/latency of the lowest number of the cheaper ram.
Click to expand...

Again, without knowing what platform it's impossible to say what will and won't work for you.

Just keep in mind that memory is binned by memory vendors to run at the capacity it's sold in. By mixing two kits together, there's still no guarantee it will work at the advertised speed.
 
Just do a couple of laps of HCI memtest, if no errors you're fine. But the consensus is still the same, there's no guarantees that the memory will run at the rated speeds (in this instance the slower kit) when doubling the density.
 
Well it's going to be a H270 motherboard with a kabylake pentium CPU the 4560.

both sets of RAM are DDR4 2400 Mhz from the same line crucial ballistix but I believe may have different timings. I will obviously run all of them at the slowest of either in terms of timings.
 
Psycho Sonny said:
Well it's going to be a H270 motherboard with a kabylake pentium CPU the 4560.

both sets of RAM are DDR4 2400 Mhz from the same line crucial ballistix but I believe may have different timings. I will obviously run all of them at the slowest of either in terms of timings.
Click to expand...

That speed, you shouldn't run into a problem on that platform. Just make sure to run a memory stress test like HCI as mentioned above.
 
I've got a related question, can you run different pairs at different timings?
For a couple of years I have run mismatched RAM but have always done so at the slower of the timings. My memory is:

Gskill Ripjaws 2*2GB PC12800 7-7-7-24
Corsair Vengeance 2*4GB PC12800 9-9-9-24

So I run these at PC12800 9-9-9-24 but in the BIOS I can set the the timings independently for each bank. So what I'm wondering is whether I can set the Ripjaws to run at CAS7 but leave the Corsair at CAS9? I've never done this because it sounded like a bit of a recipe for disaster to have timings out of sync for a shared resource like memory but maybe the system is smarter than I give it credit for and it could handle this without a reduction in stability?
 
HangTime said:
I've got a related question, can you run different pairs at different timings?
For a couple of years I have run mismatched RAM but have always done so at the slower of the timings. My memory is:

Gskill Ripjaws 2*2GB PC12800 7-7-7-24
Corsair Vengeance 2*4GB PC12800 9-9-9-24

So I run these at PC12800 9-9-9-24 but in the BIOS I can set the the timings independently for each bank. So what I'm wondering is whether I can set the Ripjaws to run at CAS7 but leave the Corsair at CAS9? I've never done this because it sounded like a bit of a recipe for disaster to have timings out of sync for a shared resource like memory but maybe the system is smarter than I give it credit for and it could handle this without a reduction in stability?
Click to expand...

from what i have been told it's the last number which is the most important iirc.

both are 24 so would make zero difference and probably a recipe for disaster.

i doubt it would make it any faster either if you could but i may be able to do the same so could you test it and report back? you will need to run a memtest though.
 
Psycho Sonny;30474687 said:
from what i have been told it's the last number which is the most important iirc.

both are 24 so would make zero difference and probably a recipe for disaster.

i doubt it would make it any faster either if you could but i may be able to do the same so could you test it and report back? you will need to run a memtest though.
Click to expand...

Where did you hear that? The tRAS window is not the most important at all, if anything it is the most flexible.
 
I have mixed and matched all kinds of different RAM before. At one point I had 2x 4GB 1600MHz Corsair Vengeance paired with 1x 8GB stick of low profile 1333MHz Patriot and 1x 8GB Samsung Green.

They were all totally different timings, and in some cases, speeds. I just manually specified my RAM settings so they were all running the same spec.

Never had an issue.
 
Acme;30489220 said:
Never had an issue.
Click to expand...

Me neither - 99% of time, you can mix it without issue, providing you set it to Auto, or manually set it to the loosest of the timings.

Generally it's only an issue if trying to run the memory at higher speeds e.g. when overclocking
 
Very true. That said though, in my case I ran all my memory at 1600MHz when two of the sticks were 1333. No surprise for the Samsung Green, slight surprise for the cheap stick of Patriot! :)
 
Unless you're looking at very high clock speeds (e.g. 3000+ for DDR4), or are trying to mix 2 dual channel sets for a quad channel system, it should be fine really. No manufacturer is going to guarantee that obviously, but you should at least be able to run JEDEC speeds and timings
 
I had a set of 2 x 4G of corsair RAM

Bought another set of identical corsair RAM

They wouldnt work together , ran ok seperately

It might work , might not
 
You must log in or register to reply here.
Back
Top Bottom