Question about memory kits stability—though...
Question about memory kits stability—though...
Hello,
I’m addressing what appears to be a common concern. I previously installed a KLEVV DDR4-3200 2x8GB package on my Ryzen 2400g system and upgraded it to 32GB. I understand the potential challenges involved. I’ve used this setup many times without any problems. This time, however, something seems different.
The system starts up smoothly and remains stable. It doesn’t crash or fail under stress tests. Occasionally, when I power it on, it briefly shuts down for a second before restarting successfully—no BIOS alerts were observed.
I’m currently running a memtest with multiple passes. If the test completes without errors, I’ll consider keeping this configuration.
Please keep in mind that this is intended for a server environment, and I sometimes spend extended periods away from it.
System details:
Motherboard: ASUS Prime B350-Plus
RAM: 4x8GB KLEVV DDR4-3200 (clocked at 3000)
CPU: Ryzen 5 2400g
PSU: Corsair RM-550x
GPUs: GT 730, GT 610, Integrated Vega Graphics (disabled)
SSDs: Corsair MP510 512GB (M.2 to PCIe adapter), Samsung Evo 128GB (boot drive)
HDDs: 5x Seagate 2.5" 320GB drives
OS: Ubuntu Server 21.04 LTS
Hey there,
Firstly, please list your full PC specs, including PSU.
Testing the mem is a good first call. It should show no errors. If there are any errors, then the ram is faulty.
How did you get a 28gb kit? That's not available for DDR4. Ideally, what you should have done was purchase a 2 x 16gb kit, and sell whatever you have to offset the cost.
What bios are you currently running on your mobo? You might need to update the bios to have ram compatibility with what you've purchased.
If I were you, I'd return the kit you bought and get a kit as mentioned above. If you are running a server, you'd want solid. stable ram, not fussing around to get it to work, when it may just start throwing up errors either way.
The shutting down you described sounds more like a PSU issue to me than a ram issue.
Server: Please revise your post to add complete system hardware details and operating system information.
What prompted the upgrade from 16GB RAM to 32GB?
Beyond the boot time shutdowns, did performance actually improve?
All RAM packages should be identical sets. I anticipate a single 4 x 8GB unit rather than two separate 2 x 8GB units, which wouldn’t be a proper match. This might not function correctly, could work perfectly, or might cause issues like those you described.
Check the servers' logs for any error codes, warnings, or informational messages recorded just before or during the shutdowns?
It could indicate another issue—perhaps an update failure, an app attempting to update, backup, or connect to a phone.
Sorry! I meant to refer to a 2x8GB kit. I purchased another one to reach a total of 32GB.
I'm not sure about that. It only works when it starts, and it functions afterward. And it happens occasionally.
Since the server is mostly active, I can't remember if it was doing this before.
It successfully passed three times. I think I'll let it pass a few more.
Sorry for not planning that earlier. It's now done!
I'm building Android. This has been something I've wanted to try for a long time. You need at least 32GB to create builds.
In fact, I'm considering getting a used Ryzen 7 3700X to significantly improve performance and speed up builds. It's an upgrade from $60-80 compared to what I've seen before.
Okay, fine. Yes, mixing ram can be problematic. Best option as mentioned is, a 2 x kit of 16gb. Rules out any potential issues. Sell what you have to offset the cost.
Well, that's your choice. However, a PSU issue can be random or not, at start up or not, or simply whilst idling or at load. Your PSU is good quality, but it could be failing. How old is it? Given it's an RMx, it's prob only3 years or so? If it's not delivering enough power or clean power it may be causing your issue.
Mem is fine. No errors after 4 passes indicates a certain amount of stability. What speed are they running at? Is xmp on?
Again, what Bios are you running on the mobo? You can use CPU-z to determine this easily.
If it begins and operates briefly on a black display before restarting autonomously and then functions properly afterward, it might indicate the motherboard is performing memory training cycles (especially if a yellow indicator appears). This behavior typically occurs once after adjusting RAM settings, though it can happen repeatedly. Regarding the CMOS battery, check your BIOS under DRAM timing controls for an option named "memory context restore"—whether it’s active or inactive.
Additionally, someone suggested using cpu-z to verify your BIOS version, but this tool isn’t available on Linux. On Ubuntu, you can use cpu-x as a substitute.
I completed more memory tests but still experience occasional restarts before POSTing. I reduced the clock speed to 2800 MHz, but it no longer works. It seems the mismatched memory chips aren't compatible at high speeds. There might be a solution, but I'm confident this is the price of the risk I took.
My BIOS is version 6042 from 2022, with several updates available. Could the latest release from late 2024 be helpful?
It looks like it's quite outdated—especially since at least 2018—and I don't have a "restore memory context" feature.