ASRock.com Homepage
Forum Home Forum Home > Technical Support > AMD Motherboards
  New Posts New Posts RSS Feed - Memory Loss after BIOS flash to 2.0?
  FAQ FAQ  Forum Search Search  Events   Register Register  Login Login

Memory Loss after BIOS flash to 2.0?

 Post Reply Post Reply
Author
Message
blppt View Drop Down
Newbie
Newbie


Joined: 29 Nov 2017
Status: Offline
Points: 17
Post Options Post Options   Thanks (0) Thanks(0)   Quote blppt Quote  Post ReplyReply Direct Link To This Post Topic: Memory Loss after BIOS flash to 2.0?
    Posted: 29 Nov 2017 at 10:43pm
Hi, new here, and I'm not sure EXACTLY when this happened, but starting from the BIOS the mobo shipped with (1.5), my X399 Taichi and TR 1950X at some point went from having almost the entire 64GB available to Windows & Linux to suddenly having only 47.9GB "usable" in both Windows 10 and Ubuntu live. If I had to guess, either the 2.0 BIOS I flashed to yesterday or a cpu microcode update caused this sudden loss of available memory. I tried downgrading to 1.5 or 1.8 BIOS (what I've been using for weeks before 2.0 came out), but either the microcode doesnt downgrade, or the BIOS was never the issue.

Its worth noting that the BIOS shows Quad-channel enabled and full 64GB. Even windows says 64.0GB installed, but only 47.9 usable.

Things I've tried:

1. Loading UEFI defaults in the BIOS
2. Disable/Enable XMP
3. Bumping SoC Voltage to 1.1 from 0.9 default.
4. Bumping LLC to Level 1
5. Loading the latest Ubuntu (17.10) from a live USB stick, still only 47GB available.
6. Swapping the sticks from the recommended (2) slots to the (1) slots for troubleshooting.
7. Running a memory test with windows 10 memory checker, no errors.
8. Downgrading to both 1.5 and 1.8 bioses, no change---not sure if this goes to an earlier version of cpu microcode or not.
9. Taking the 4 Corsair Dominator sticks from my other box and swapping them in, same result
10. Changing from NUMA to UMA mode, no change.

Like I said, at some point after flashing from 1.5 to 1.8 (the day I set it up, weeks ago) the RAM was working the way it was supposed to in Windows, with nearly the entire 64GB available (63.9 or something like that), but I just noticed after flashing to 2.0 yesterday and running benches that only 47.9 was available.

The Kingston Ram I use for this box is on the approved list, 4 sticks: HX426C16FBK2/32

The Corsair Ram I troubleshooted with is CMK64GX4M4A2666C16, 4 sticks.

Anybody have any ideas?
Back to Top
MisterJ View Drop Down
Senior Member
Senior Member


Joined: 19 Apr 2017
Status: Offline
Points: 1097
Post Options Post Options   Thanks (1) Thanks(1)   Quote MisterJ Quote  Post ReplyReply Direct Link To This Post Posted: 29 Nov 2017 at 11:57pm
blppt, hopefully the memory experts will comment, but I suspect you have a bad (weak) stick of memory or one that is not properly seated.  So first do an F9, F10 after re-seating all your memory sticks and see if all 64 GB appears in your OS.  If you are short a stick, then remove 3 sticks and boot using one stick at a time in the same socket.  If all boot, then perhaps you have a bad socket.  Test each socket, one at a time, with one stick.  I assume you are using A2, B2, C2 and D2.  You could also try A1, B1 etc.  I had to RMA a X370 board with bad A1 and A2.  Please let us hear what you learn.  Enjoy, John.
Fat1 X399 Pro Gaming, TR 1950X, RAID0 3xSamsung SSD 960 EVO, G.SKILL FlareX F4-3200C14Q-32GFX, Win 10 x64 Pro, Enermx Platimax 850, Enermx Liqtech TR4 CPU Cooler, Radeon RX580, BIOS 2.00, 2xHDDs WD
Back to Top
blppt View Drop Down
Newbie
Newbie


Joined: 29 Nov 2017
Status: Offline
Points: 17
Post Options Post Options   Thanks (0) Thanks(0)   Quote blppt Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 12:07am
Yeah, I tried all that---including trying 4 completely different sticks of DDR4 from another box and the A1, B1, etc versus the 2s. I just think its weird that the 2.0 BIOS flash yesterday was the first time I noticed the RAM "available" level drop---plus, wouldnt the RAM not show up in the BIOS at all if they werent seated properly? The BIOS shows all 4 sticks, 16384, and Quadruple Channels.

I was hoping it was just a bug with the latest CPU microcode, but I'm starting to think maybe its a bad memory channel; the problem is obviously not with the RAM itself, as the exact same issue shows up with the memory from my other box, which is an Intel setup with Corsair memory with the full 64GB available. But, put that very same memory in my X399/AMD system, and the 47.9 shows up.
Back to Top
jmmec View Drop Down
Newbie
Newbie


Joined: 22 Nov 2017
Status: Offline
Points: 21
Post Options Post Options   Thanks (1) Thanks(1)   Quote jmmec Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 12:50am
Seems others are having similar problems, but on an older BIOS - check out this post in particular?

http://forum.asrock.com/forum_posts.asp?TID=6546&PID=40043&title=x399-taichi-boot-loop-after-flashing-to-bios-v180#40043


Back to Top
blppt View Drop Down
Newbie
Newbie


Joined: 29 Nov 2017
Status: Offline
Points: 17
Post Options Post Options   Thanks (0) Thanks(0)   Quote blppt Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 1:22am
Yeah, I saw that post (and others) earlier, the only thing I havent tried that they did was re-seat the CPU (suggested in another thread and actually solved his problem) , which I guess I will try later today when I get home from work, but I am not optimistic that is going to solve anything---after all, the memory *did* work just fine at some point, and I havent done anything to the water cooler mount to make me think the cpu somehow shifted.
Back to Top
MisterJ View Drop Down
Senior Member
Senior Member


Joined: 19 Apr 2017
Status: Offline
Points: 1097
Post Options Post Options   Thanks (1) Thanks(1)   Quote MisterJ Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 4:31am
blppt, I have seen not completely seated memory be detected in BIOS but not be available in Windows.  I think the memory (perhaps channel) fails POST.  Did you test all 8 of your memory sockets?  Please let us know what you learn.  Enjoy, John.
Fat1 X399 Pro Gaming, TR 1950X, RAID0 3xSamsung SSD 960 EVO, G.SKILL FlareX F4-3200C14Q-32GFX, Win 10 x64 Pro, Enermx Platimax 850, Enermx Liqtech TR4 CPU Cooler, Radeon RX580, BIOS 2.00, 2xHDDs WD
Back to Top
blppt View Drop Down
Newbie
Newbie


Joined: 29 Nov 2017
Status: Offline
Points: 17
Post Options Post Options   Thanks (0) Thanks(0)   Quote blppt Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 5:08am
Nor have I---but it appears perfectly in the BIOS, and technically Windows detects that 64GB is installed, but for some reason, only 47.9GB is usable. I'd say it was a Windows issue except the exact same thing occurs booting into Ubuntu 17.10 via the live stick.

I've tested the A1,B1, etc. slots, and the A2, B2 slots with all 4 sticks, and the same thing happens. Its almost like theres *some* option turned on the BIOS for an onboard peripheral thats causing ~16GB worth of memory addresses to be taken up for no reason, but I'll be darned if I can figure out which one that is. I've already tried moving my video card (Vega 64 w/c) from the 2nd slot to the 1st slot, no changes. Gonna have to try reseating the cpu tonight, but again, I dont expect that to change anything.
Back to Top
MisterJ View Drop Down
Senior Member
Senior Member


Joined: 19 Apr 2017
Status: Offline
Points: 1097
Post Options Post Options   Thanks (1) Thanks(1)   Quote MisterJ Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 6:22am
blppt, many have seen this before, including me. I do not think it is a BIOS issue and certainly not an OS issue.  I suspect that the BIOS tests the memory and when all but one stick is good, passes that information to the OS.  Have you tested each slot, one at a time with one stick?  You may find a pair (A1 and A2, like me) that is bad.  I had to RMA my board.  Good luck and enjoy, John.
Fat1 X399 Pro Gaming, TR 1950X, RAID0 3xSamsung SSD 960 EVO, G.SKILL FlareX F4-3200C14Q-32GFX, Win 10 x64 Pro, Enermx Platimax 850, Enermx Liqtech TR4 CPU Cooler, Radeon RX580, BIOS 2.00, 2xHDDs WD
Back to Top
blppt View Drop Down
Newbie
Newbie


Joined: 29 Nov 2017
Status: Offline
Points: 17
Post Options Post Options   Thanks (0) Thanks(0)   Quote blppt Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 7:08am
I feel like such an idiot....the bottom two screws on the cpu loader mechanism were either never tightened, or somehow came loose. As soon as I tightened them up, the problem was fixed.

So, moral of the story is, sometimes "reseating" the cpu  (in quotes because it was probably never seated properly in the first place) CAN solve inexplicable problems. Personally, I'm shocked that it was rock stable even with the CPU not making perfect contact with the pins. Wacko
Back to Top
MisterJ View Drop Down
Senior Member
Senior Member


Joined: 19 Apr 2017
Status: Offline
Points: 1097
Post Options Post Options   Thanks (1) Thanks(1)   Quote MisterJ Quote  Post ReplyReply Direct Link To This Post Posted: 30 Nov 2017 at 8:12am
Great, blppt!  Thanks much for reporting.  Enjoy, John.
Fat1 X399 Pro Gaming, TR 1950X, RAID0 3xSamsung SSD 960 EVO, G.SKILL FlareX F4-3200C14Q-32GFX, Win 10 x64 Pro, Enermx Platimax 850, Enermx Liqtech TR4 CPU Cooler, Radeon RX580, BIOS 2.00, 2xHDDs WD
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.04
Copyright ©2001-2021 Web Wiz Ltd.

This page was generated in 0.156 seconds.