942-Memory Training error (code 3016)

[gghiddi]

Hi to all, sorry for my basic english, i have buied a workstation z620 dual cpu with dual xeon e5-2670, 64 gb ram, ssd and gtx-1070 refurbished from a professional ebay reseller.
Always, on power on the workstation, before the s.o. boot and startup, the system give me the message :
"942-Memory training Error. Dimm 3 on cpu1 experienced an error during training (code 3016)
F1 : boot", pressing f1 the workstation boot fine. i' ve switched the ram from cpu0 board and cpu1 board and the message is the same, i ve used the memory test on bios of the workstation and no error is given, i ve tested ram with memtest86+ for 14 hours and no error is given.
on windows 7 x64 the memory is all 64gb and the workstation seems to work fine.

somebody may help me to solve problem ?

Hi and thank's to all for reply and help.

have a nice day.

[joselso]

The 942-Memory Training Errors, with various codes (1501, 301C, 3016), have been encountered in several HP Z-series workstations - Z420, Z620 and (my own) Z820, since the introduction of Intel's family of microprocessors early in 2012.

This type of error is usually accompanied by a large number of repeated WHEA (Windows Hardware Errors) Warning Events in Windows 7 Event Viewer:

Event ID 47 - Memory, Corrected Machine Check and Event ID 20 - Microsoft-Windows-Kernel-WHEA/Errors.

I recommend that you verify it in the Event Viewer in your Z620 after booting Windows 7.

No 'Memory Diagnostics' could find the error; but it kept appearing at every booting.

There are two 'workarounds' that I tried:

1. Similar to recommendation in HP Customer Advisory: URL:

http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c03282091  :

In the BIOS Setup (F10) of the Z620 Workstation, in 'OS Power Management', change 'Idle Power Savings' from 'Extended' to 'Normal', and see that 'S5 Maximum Power Savings' is set to 'Disabled'; by this I managed to reduce the rate of WHEA warnings in Windows 7 Event Viewer from thousands per second down to a few dozen.

Yet, this workaround alone still did not prevent the '942 - Memory Training Error' on every booting of my HP Z820.

2. Having moved the suspected Memory DIMM among slots and seeing that the location of the 942 error followed the location of the DIMM, I informed HP Support and their technician verified my findings and replaced the faulty Memory DIMM with a new one. (I have purchased extended warranty for my Z820).

A similar chain of events occurred in my Z820 workstation several times between 2013 and 2016,
and I expect to encouner it again during the prospected life of the workstation.

As your Z620 is now out of warranty, you will probably have to experiment with various types of replacement Memory DIMMs before the '942 - Memory Training Error' goes away for a few months or more.

[joselso]

Note: in case the shortcut

http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c03282091

 to HP  SUPPORT COMMUNICATION - CUSTOMER ADVISORY fails to reach the HP document, you can reach it if you search Google for "Windows WHEA Logger Logs Frequent Event ID 47 Error" .

[gghiddi]

hello, I 've checked the Event Viewer on Windows 7 Filtering for id 20,47

for all the registers, no errors found for ID 47 and ID 20 only for other causes and until 04/11/2016,

I also tried swapping the ram on more slots, but it does not work, the error is always given on "DIMM 3 oN CPU1", but in this slot I tried 3, 4 different memory bank.

so, the warranty hp definitely is ended but the ebay vendor warranty is one year and are still in the period of "withdrawal from purchase" I have to know if the defect is severe and then return the workstation to the supplier or request a replacement.

Have a nice day 🙂

[joselso]

If the error stays at the same slot and does not follow the memory DIMM, then probably there is some fault at that slot or in devices to which it is connected.

You may follow the advices given by Dan_WGBU on a similar case, discussed in:

http://h30434.www3.hp.com/t5/Business-PCs-Workstations-and-Point-of-Sale-Systems/Error-on-689471-001...

[erichfrazer · Accepted solution]

This was my problem: I had a bent pin on the CPU socket. I got myself a thin needle and a magnifying glass and successfully bent the pin back into place, but almost wasn't able to. I had to remove the entire mobo from the unit in order to focus on the bent pin.

the clue is if the error doesn't follow the chip, if it stays no matter what chip you put into it, then the problem isn't the chip.

[gghiddi]

Thanks, i have already sent the workstation to the supplier for a check since the solutions he suggested did not have results and that could be a pin bent, i have already got back the workstation and no more error

[0x87f22]

I thought i'd contribute to this.

I had a Z820 with this training error.

Turns out the indicated module had an extremely tiny chip on one of the components - presumably a resistor or capacitor. It was literally a fraction of a milimeter - barely visible - but it was enough to throw off the memory timings intermittently.

The chip was located close to the top edge of the module, so pressure when pushing it into the DIMM slot would've caused some stress on this component.

So, the problem (for me) was big fingers and a poorly designed memory module which has delicate components located a little too close to the top  🙂