I’m aware of memtest86 / occt / memtest64 / Ryzen dram calc’s built in tester and my setup passes all of these fine for sustained durations. However, I saw a YouTube video in the past where they found that their ram only became unstable during long sustained gaming sessions when their gpu was under load and the whole system was generating lots of heat. My understanding is if you’re memory is right on the edge of stability it might be throwing errors here and there during games / causing stutters or some such as a result and you wouldn’t even know it since unlike these mem testers they don’t throw visible ram errors for you. Anyhow memtest86 seems to be the gold standard and I’ve passed that so maybe I’m just being paranoid. If ram is unstable during gaming are their specific tell tale signs or not really? Thanks,
How would you point to ram while gaming when "whole system is generating lots of heat"? . In the history of PC gaming RAM was never an issue until Ryzen came out. I'm yet to see a single Intel system user asking question about RAM... while Ryzen user has stuttering systems, changes ram settings to find a temporary solution. Only later to start fixing it again because of most likely some overheat. If you are really paranoid maybe it's time to switch.
ignoring what Freeman has said as there is no useful information at all.... Ryzen users like to tweak ram performance as it can improve CPU and system performance. Otherwise, for sake of stability, can often leave ram at default xmp or jedec speeds. I've had it occur where my ram after tweaking seemed stable but later turned out not to be. This is one of those things with ram and not necessarily due to gaming, it can be memtest stable for 24 hours and then turn out it's not later. For reference, I had these difficulties tweaking when I had Hynix JJR modules, they weren't the best sticks of ram, I now have Hynix DJR modules which are much much better. For me, much like you figured, once the system reached a state where it had been heated due to full load, ram became unstable as the ram tweaks were only stable after a cold boot. Tell tale signs of ram instability would be random program crashes, strange behaviour in programs too or otherwise full on system crashes. If you find heat is a contributing factor to instability of your ram, you'll need to dial voltage down a little (and if necessary loosen any timings or reduce any overclock). Tldr is though, memory error hunting can be a difficult process, ram can be memtest stable after a long time just to have an error slow up later, just be vigilant in your observations.
To clarify my system isn’t overheating, neither was theirs in the video to my recollection, I should’ve said memory instability seemed to only occur when the other components were under load and wasn’t caught by mem tests for them. In the comments I recall many people mentioned heat being necessary and/or gpu load being required for the instability to occur hence my question. I’ve been trying to dig up the video, but I won’t be able to look more thoroughly til after work. Also fast RAM is not a temporary solution, Ryzen really is sensitive to memory speeds and timings, am I misunderstanding your meaning? Thanks, But yes, I hear you — in the future Intel not requiring super fast RAM for max perf may be relevant to my purchasing decision. If I’m misinformed on this whole thing that’s fine too, just trying to ensure system stability as best I’m able.
That's not strictly true anymore, Rocket Lake has proven to benefit from fast timings just like Ryzen does.
Thank you very much, this is helpful! I hear you, for me I would just run the system with XMP defaults but my system won't boot with those so I've had to manually test what works and what doesn't. According to memtest86 for example I can achieve significantly lower primary timings albeit with a bit less frequency at the recommended voltage. It's notable that occt / ryzen dram calc did not catch errors that memtest86 and memtest64 did catch hence why I've tried all 4 (memtest86 seems the most thorough). As far as issues in windows or games, I haven't had any crashes or failed boots and temps look fine during gaming, but I ran into a couple long frames here and there during a long gaming session that made me wonder what the tell tale signs of ram instability would be hence this post to see if there were more ways to test.
Tooooooooooooooooooooooooooooooooooooooooo many variables... But some memory chips are more sensible to temperature, keep a good airflow and you should be fine. Wrong voltage settings(VDDG IOD and mainly SOC) may cause a problem when the GPU/NVMe SSD/USB are working at the same time. You may try to copy a BIG file during a game and check the copied file for CRC errors.
Thanks! That's helpful -- in your experience if the memory passes memtest86's full test suite am I probably good to go? My experience thus far has been memtest86 / memtest64 were able to catch an error that occt / ryzen dram calc's memtests did not interestingly. Good idea with the file copy, will try that out.
Any particular recommendation? If you can’t be sure of RAM stability with the best memory testers then I’m not sure how you can ever be sure without running stock settings. If XMP fails to boot then the only way forward is manual painstaking testing and use of these testers no? Thanks,
You can't prove the absence of a problem, there will always be the chance there Equally you can't prove something isn't the source of an issue without isolating the component.
I am glad to help, your mileage will vary, but try http://www.numberworld.org/y-cruncher/ (press 1-7-0)
