I have a new (refurbished) Z620 with a Quadro M4000 8gb, 32 gb RAM, Windows 10. The processor is an Intel Xeon E5-2680@2.7 GHz, 8 Cores.
The BIOS is HP J61 v3.90, 6/1/2016.
I noticed that when I left the computer for a while, I’d come back to find a black screen with this error:
929 Fatal MCA Error; MLC Error Detected CPU 0. Internal timeout error Watchdog Timer (3 strike).
Reading about what this could possibly be is a little scary; got all set to run memory diagnostics and re-seat the proc (or send it back under warranty), but I noticed that the crash only happened when I left for a while.
I checked the power settings, and saw that «Sleep» mode was enabled, timed for 30 minutes or so of no activity. I changed this to 5 minutes and hung out a while, and it crashed with the same message. I changed it to «Never», and it has been on ever since (two days and counting).
In my case, I don’t think it is a hardware error at all. I did run some memory diagnostics (Windows) and found no errors. I will boot from a USB and run MemTest overnight to be sure — but I think the hardware is fine.
From what I have read, this is sometimes a problem of power reaching the CPU or the RAM; if it happens when the computer goes to sleep, that’d make sense.
I never use these automatic sleep settings anyway, so it’s not really an issue except that it’s wrong and it bugs me — anyone have any thoughts about it?
If my MemTest shows errors, that’s that — if not, I will pull chips and test with the 5-minute sleep test. I think it’s a BIOS setting — would love to hear from someone with more experience than I.
With Win 7 Pro 64 bit previously, and now with Win 10 Pro 64 bit, 929 Fatal MCA error QPI0, 929 Fatal MCA error QPI1, and 932 Warning one of the QPI links is not operating, have appeared, and in the past reseating the CPU1 riser has temporarily solved the problem until recently! I note that this problem has been archived and have read that post!
In my case reseating the CPU1 riser no longer worked!
Close inspection of the 8 conductor on the motherboard revealed deterioration of the plastic on at least 2 of the contacts! This probably resulted in inadequate contact with the CPU1 riser, resulting in the 929 and 932 errors! Repair of the motherboard conductor contacts is impossible from within the case! Solution might be possible if I could hold the connector close in my hand! After considerable searching to find how to remove the right hand panel (HP Service Media Library D/C Nov 2019) I eventuall figured out how to remove the right panel, which allowed me access to the connector from the motherboard to the CPU1 riser! With the connector in hand I attempted to repair the 2 contacts which appeared not to align properly! After reassembling the connector and replacing it in it’s proper position, I replaced the right hand panel after vacuuming the case!
With great relief I reassembled the CPU1 riser to the Motherboard, reattached the cables, and presto, successful boot, and thought that while this currently solved my error problems, that perhaps this information may be of some value to others with a Z620 and dual Xeons!
Thank you!
I just re-seated the CPU this morning. I cleaned all the liquid metal off the surface and went back to regular thermal paste for now. No difference, it is still crashing on startup (gets to the welcome screen but then like clockwork, it freezes there if I have hyperthreading enabled. After it crashes, the system will reboot and I will get this error (from the z840 system alerting): If I disable hyperthreading the system passes any and all benchmarks and boots into windows without a hitch. I can also run prime without any crashing going on.
929-Fatal MCA error
Core 0 generic L2 cache generic error — poisoned data
CPU: 1
Status: 0xFE20000000000C110A
Misc: 0x0000000000000000
Address: 0x0000000000000000
Press ENTER to continue
People say this could be a memory issue manifesting itself as a CPU problem, but I have spare memory and swapped that in and it made no difference. Reason I am thinking its more CPU related and not memory is because the initial problem happened when I started running an AIDA64 torture test. it crashed under that load and I was never able to use hyperthreading again.
Not the end of the world I still have 36 cores to throw into the mix lol
Also, I can upgrade to v4 processors at some point if I can narrow this problem down and say definitively it needs a CPU.
I suppose the next step, since this is a dual processor rig, is to swap processors and see if the error gets tripped in the CPU 0 socket. Any other troubleshooting suggestions here?
Here is a look at eventvwr.msc and I don’t know if it’s related but just though I would iiclude a snip…
See less
See more
Pretty sure it’s the CPU guys…
dayum dude, I got a Z440 mobo in a 3u server chassis with a 2660 v3, I wish I’d had seen this chip though… I hope you get it figured out.
You mentioned a dual CPU rig, so definitely run the machine with either CPU and see how far you get. Those Mobos aren’t expensive to replace if it’s a problem, thank god.
- This is an older thread, you may not receive a response, and could
be reviving an old thread. Please consider creating a new thread.
-
#1
So, I have reason to believe that my existing Xeon E5-1650 V3 CPU is faulty. I have recently swapped it into a new board and I, after installing Windows 7 on it, the workstation (A HP Z440) reboots and I get a message saying:
«929-Fatal MCA error
CBO/LLC2 error detected on CPU 0
Core 0 Data L2 cache data read error
CPU: 0
Status: 0xFE20000000041136
MIisc: 0xA4F00070E0001086
Address: 0x00000000CF8E1680″
I’m located in Australia but if you happen to have one of these chips and don’t need one and are willing to offload it for a reasonable price, let me know.
Screen message
922—Front USB 2 Not
Connected
923—Fatal IRPP error.
924—Fatal IIO error
925—Fatal Misc. Error
927—Fatal error on DIMM in
slot CPU X DIMM Y
928—Fatal error occurred in the
designated slot.
929—Fatal MCA Errorr
939—Front USB 3.0 Not
Connected
940—Front 1394 Not
Connected
941—PCIe Device(s) installed in
slots 3 or 4 with a single CPU
942—Memory Train Error
1801—Microcode Update Error
1802—Processor Not
Supported
Probable cause
Front USB 2 mechanism is missing or is
not connected.
A fatal miscellaneous chipset error is
setected.
Fatal multibit ECC error detected on the
DIMM in the slot labeled DIMM Y (where Y
is a number), as labeled on the system
board.
Fatal error occurred in the designated
slot.
An MCA condition is detected on the
system.
Front USB 3.0 mechanism is missing or is
not connected.
Front 1394 mechanism is missing or is
not connected.
Invalid system configuration.
A DIMM or DIMMs did not train correctly.
Unknown or unsupported processor
stepping.
The system board does not support the
processor.
Recommended action
2.
Reseat or replace the missing mechanism cable.
1.
Replace or reconnect the missing Front USB 2
mechanism to the system board.
2.
Reseat or replace the missing mechanism cable.
Potential system problem; contact HP Support.
Potential system problem; contact HP Support.
Replace the DIMM in the identified slot.
Move the card to a different slot. If the problem persists,
replace the card.
1.
Replace or reconnect the missing Front USB 3.0
mechanism to the system board.
2.
Reseat or replace the missing mechanism cable.
1.
Replace or reconnect the missing Front 1394
mechanism to the system board.
2.
Reseat or replace the missing mechanism cable.
1.
Move the PCIe device from PCIe slots 3 and/or 4 to
other PCIe slots.
2.
Add a second processor using appropriate HP
accessory kit.
The microcode update failed. If the processor stepping is
supported, contact HP Support.
Replace the processor with a compatible one.
Diagnostic codes and errors
97