Windows only detecting 16GB of 32GB

Geno750

Gawd
Joined
Apr 28, 2006
Messages
688
Noticed the PC was starting to act strange in solid works, almost like it was out of ram. I'm using an X99 mobo with 32GB of ram installed. The bios sees it, CPU-Z sees it, but windows doesn't. Is this a windows issue or should I be looking at trying a beta bios from Gigabyte (already on latest non-beta bios).

kFuTf5C.jpg


Kt2LFe2.jpg


ouwttdN.jpg
 
Was it working before? Step one would obviously be to reseat all the RAM. After that, consider a beta BIOS. If that doesn't work, run memtest on it overnight and if you get errors RMA the RAM. Have access to other sticks you could try? Is the RAM you're using on your motherboard's QLV list?
 
Did a Mem test and came up good, going to move them into the other banks and see what happens.
 
Go into MSCONFIG and under Boot tab, check Advanced options. Is the maximum memory set to 16GB?
 
Windows 10 Edu can handle a crap load more than 16Gb thats for sure...

Can you try another OS to see if it is just windows? maybe a linux live boot cd?
 
Are you overclocking your RAM? Try running stock to see if that helps. So many issues with overclocking memory on X99 that it pisses me off just thinking about the hours spent troubleshooting it...
 
Ram isn't oc'd. It's rated to run at 3200mhz. However I've dropped it down to 2800 to troubleshoot with. I've been messing with it most of the day. Moving the sticks around suddenly made the system very unstable. Finally tried loading just the right 4 banks with all four sticks and it booted and loaded Mem test. The test is going on now and I'll let it run overnight. It's now looking more like a possible motherboard issue.
 
I vote for trying a 64-bit live boot disc of a Linux distro & verifying the max RAM available. That will tell you if Windows is having issues.
 
If your motherboard has a ram mirroring option, check that it's not enabled. It's a server feature that makes your ram act like raid 1 hdds. Including getting half the capacity as a result.
 
I'd boot Linux off a flash drive and if ram works OK then it's not a hardware problem but a flaky OS.
 
So after a lot of testing, it seems its most likely a CPU issue (possible mobo but doubt it now). When operating as quad channel, at 2666MHz memory bus I get all 32GB of ram showing in system. Things get fun when I go to 2800MHz or greater, at those speeds I'll get 16GB (only right banks of memory are active). 3200MHz I get either 16GB or sometimes 8GB. If I populate only the right banks on the motherboard, I can get 32GB at 3200MHz. This is at any CPU speed, even underclocking the CPU nets the same result. I can only assume my IMC is no longer happy at anything greater than 2666MHz. The CPU is still rock solid at 4.5GHz, and I've even been able to tighten up the memory timings with the new memory bus limit. I may try a different set of ram out to completely rule out memory as an issue. However, I've been running memtest in quad channel mode at the new fsb and there were no issues.

bSva98k.jpg


B1iPF35.jpg
 
That's an unfortunate (but based on my understanding of how these things work, likely accurate) diagnosis.

Yeah, other than testing with another set of RAM, seems you've narrowed it down. Mobo's still a possibility, but IMC seems more likely.
 
What is the motherboard you're using? I don't think it's a CPU issue since I've had experience with MSI motherboards not liking the G.Skill RAM I used. The MSI mobo wound up dying, so I go it RMA'ed & bought an ASUS mobo instead. Now my RAM can run with the XMP settings just fine.

Either update the UEFI or get it RMA'ed if you're still having issues.

Edit:I see you have an ASUS X99 TUF... so that's really odd. I'm using Ripjaws 4, so I don't know how well the TridentZ series works with X99.
 
Actually this is in a gigabyte x99 Phoenix mobo. I may grab another one from microcenter and see what happens. The RAM is on gigabyte's list of compatible ram so it's odd that it'd suddenly stop working.
 
Actually this is in a gigabyte x99 Phoenix mobo. I may grab another one from microcenter and see what happens. The RAM is on gigabyte's list of compatible ram so it's odd that it'd suddenly stop working.

Is your signature out of date? I'm not a fan of Gigabyte, so I'm cautious about their reliability & compatibility with RAM. I've learned hard lessons with Gigabyte & MSI. So far ASUS has provided me with the best experiences.
 
Asus fucked me on a warranty (that sabertooth mobo) so I try to avoid them at any cost. The sabertooth had a bad second PCI-E slot. The RMA return I received was a used mobo with bent pins on the CPU socket, which Asus says I caused.
 
Wasn't this a known thing with X99 (maybe it was X58)?

Seems to me I even remember Kyle fighting with this and doing an article on it.

I also remember reports of people RMAing several brand new mobos before getting one that worked.

Then like in this case, it was reported to be a moving target kind of a problem with it also being related
more to the particular CPU. May have been CPU socket related, dunno.

I don't remember the specifics or what if any solutions though.
Hopefully you can dig up info on this and narrow it down.

Found it:

https://hardforum.com/threads/ram-issues-on-x58-boards.1546247/

.
 
It appears that X99 has similar issues since not all of them work the same with different RAM configurations & XMP profiles.
 
Ram isn't oc'd. It's rated to run at 3200mhz. However I've dropped it down to 2800 to troubleshoot with. I've been messing with it most of the day. Moving the sticks around suddenly made the system very unstable. Finally tried loading just the right 4 banks with all four sticks and it booted and loaded Mem test. The test is going on now and I'll let it run overnight. It's now looking more like a possible motherboard issue.

2133 is stock speed for X99, no matter what the ram is "rated" for. That you can run at 2666 but drop ram at 2800 is the clue. As others have mentioned, OCing (and even 2666 is OCd) ram on X99 can be a bit more challenging.
 
2133 is stock speed for X99, no matter what the ram is "rated" for. That you can run at 2666 but drop ram at 2800 is the clue. As others have mentioned, OCing (and even 2666 is OCd) ram on X99 can be a bit more challenging.

True, however it's been working for several months now, and has slowly been declining in capability. The randomly not working points to something not related to the oc being the issue.
 
It seems to depend more on luck than anything for my rig.

Seeing as the benefits of OCing DDR4 are miniscule at best outside of benchmarking, I run my RAM at 2133 and overclock everything else. Unless you actually enjoy spending more time restarting/tweaking OC and dealing with system hangs than using your PC...

I do enjoy it. But when you finally think things are stable only to freeze up randomly later on (or get an incorrect total size reported upon the next boot), it becomes more trouble than it's worth for absolutely no real world gains.
 
I built this system back when the X99 TUF was released and had it running at the settings in the first post. RMA'd the mobo when I realized it had a dead second PCI-E slot (second 980ti to play with SLI, which I hated). Switched to the gigabyte board I have now, which ran the exact same settings just fine until a few days ago. The sudden degradation made me curious to it being software since I figured hardware would work or it wouldn't. The system has continued to degrade, where 2666 no longer works. I'll try a new mobo this weekend and see what happens.
 
that's OK windows 10 cannot detect that my CPU is running at 4.3Ghz as my multiplier is set to 43 :eek:

oh, well it still screams when you game...
 
True, however it's been working for several months now, and has slowly been declining in capability. The randomly not working points to something not related to the oc being the issue.

I was fighting with a similar issue. I had a gaming box that held a certain level of OC for two years until it became unstable. I had to drop the clocks more and more and finally it refused to boot even on stock setings.

I was sure I fried the cpu/ram/mobo but no... It was the PSU. After I replaced the old PSU the system overclocked better than new.
 
I was fighting with a similar issue. I had a gaming box that held a certain level of OC for two years until it became unstable. I had to drop the clocks more and more and finally it refused to boot even on stock setings.

I was sure I fried the cpu/ram/mobo but no... It was the PSU. After I replaced the old PSU the system overclocked better than new.

Just finished moving the system to a new motherboard, and it has the exact same issues. So I'll have to give this a shot and see what happens.
 
Back
Top