• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Memtest kills Windows 10

G

gkfr876t39

New Member
Joined
Aug 28, 2019
Messages
5 (0.00/day)
So, I've been testing my GeiL 3000 AMD DDR4 memory - XMP profile.
Run Memtest64, everything freezes, system unresponsive. I try to reboot, Windows cannot boot.
I try repair disk, cannot fix boot problems. Have to re-install Windows. Sigh.

Then I do the same test on freshly installed system. Run Memtest64, it crashes PC, cannot boot, has to re-install Windows.

Third time I turn off XMP profile, Memtest64 runs with no issues, no errors.

I kind of understand how overclocked memory may cause PC to crash, but I don't understand why it (or Memtest64?) crashes it to the point that it cannot boot????
 
Sometimes the issue is xmp raises the BLCK too high and it crashes, Other times you need to set timmings manually to the factory clock timmings for X.M.P to work.
Instead of using X.M.P. try setting your ram manually, if it's still giving you trouble raise the Dram voltage a little along with the vccio and cpu system agent voltages and see if that gets you stabilized, raise each voltage in increments of .01v
Errors after enabling X.M.P can also be related to faulty PSU. Strange problems offen are.
 
Last edited:
Eskimonster said:
Sometimes the issue is xmp raises the BLCK too high and it crashes, Other times you need to set timmings manually to the factory clock timmings for X.M.P to work.
Instead of using X.M.P. try setting your ram manually, if it's still giving you trouble raise the Dram voltage a little along with the vccio and cpu system agent voltages and see if that gets you stabilized, raise each voltage in increments of .01v
Errors after enabling X.M.P can also be related to faulty PSU. Strange problems offen are.
Click to expand...
I would play with settings if I didn't have to reinstall Windows after every attempt. It does not make sense to me that it causes boot failure, but still POSTs.
 
you dont have other ram kits laying around do you ?
Also my PSU is giving up, i cant OC my RAM at all or raise my CPU clocks without crashing, im buying new PSU on friday :)
Even got 2 horizontal streaks on my screen lol.
 
Make a USB boot drive with memtest86 on it, and avoid booting windows until it is stable. Last time I did this, I used memtest86+ (note: almost same name), but it hasn't been updated in years, so I would try memtest86 first, I have no idea how good it is though :)
 
Memtest86 is an excellent tool for diagnosing memory issues. In my experience, NMVe SSDs can have all kinds of issues from changes to BCLOCK. I'm not sure XMP does that... XMP doesn't always work right with Ryzen CPUs. I have a Corsair kit that will run better than XMP settings but crashes with XMP on both Ryzen 1700X on X370 and 3800X on X570. I have a G.Skill kit that works perfectly with both platforms on XMP. Makes no sense...
 
As you guys recommended - I've tried memtest86 8.x (fast, but limited to 4 passes in free version), and old version of memtest86 4.x (slower, but allows unlimited passes).
Both tests found no errors.
Which made me think if memory is an issue at all, and if not, what else it could be? The difference between memtest86 and memtest64 is that the latter swaps heaps of memory onto disk. So that made me think about SSD. I have replaced the SATA cable and re-inserted the RAM modules - with firm fit.
Guess what?
More testing is due, but all good and stable now. Memtest64 passes with XMP on. And my SSD is twice as fast as it was before.
So I should change the topic of this thread to "Bad physical connections kill Windows".
 
Its an old program(junk!), why were you testing memory to begin with?
 
gkfr876t39 said:
Memtest86 version 8.2 was released in June 2019. New and not junk :)
I had frequent intermittent freezes (5-20 seconds), which made me doubt RAM.
Click to expand...

Glad you got around to running v8.2, I always test with this first b4 booting into windows or any OS for that matter when experimenting with new ram speeds & timings.
Personally I run at least 2 passes but default is 4. Imo it depends how obsesseive one is with stability if run more than 2 passes.
 
gkfr876t39 said:
So, I've been testing my GeiL 3000 AMD DDR4 memory - XMP profile.
Run Memtest64, everything freezes, system unresponsive. I try to reboot, Windows cannot boot.
I try repair disk, cannot fix boot problems. Have to re-install Windows. Sigh.

Then I do the same test on freshly installed system. Run Memtest64, it crashes PC, cannot boot, has to re-install Windows.

Third time I turn off XMP profile, Memtest64 runs with no issues, no errors.

I kind of understand how overclocked memory may cause PC to crash, but I don't understand why it (or Memtest64?) crashes it to the point that it cannot boot????
Click to expand...

Bump memory volt to 1.35, but try not to increase SOC voltage.
 
Another idea...build a WinPE boot drive, and put windows based Memtest64 on it? Could even burn it to a CD/DVD to avoid it getting corrupt
 
If memtest is freezing (I tried it on mine it's unresponsive until it passes) it has to be a ram issue, either timings or voltage is my guess. Try 1.35v like shrimp said (it's the default for the geil ram I found) and if you can't use xmp, manually set the first few timings (the 18-19-whatever, leave the rest auto) and if it doesn't work at 1t command rate try 2t with same timings.
 
You must log in or register to reply here.
Back
Top