Troubleshooting crashes and instability of my system

[Teemsan]

I've been having system crashes and BSOD's with my system. It started when I installed a new mobo - an ASUS Z87 Plus. I did RMA the first mobo thinking it was suspect, but the problem persisted. I should say from the outset that my RAM is not on the QVL, so that could be the problem. I did test the RAM with Memtest and it passed. It was also in another system that ran fine.

What happens is that my system will either freeze or I'll get a BSOD - sometimes it correlates to exceeding memory usage - which I have done, and sometimes randomly like when I'm watching a vid or just web surfing.

Two of the BSOD's were:
BAD_SPOOL_HEADER
and
REFERENCE_BY_POINTER

There have been others but the system rebooted in those cases before I could get them.

Last night it was just a freeze up. My mouse and keyboard froze, Ctrl+Alt+Del did nothing. I was just web surfing. I had to reset by power. These problems occur about once a week or every two weeks.

Initially the problem was happening frequently, but someone suggested DL'ing the latest chipset drivers for 4th Gen proc's from ASUS. That helped for sure. But the problem still happens, just less frequently.

The system:
i7 4790 (non K, so no OC)
ASUS Z87 Plus mobo
GSKill Ripjaws X Series 16GB (2x8) F3-1600C9D - I have two of these kits installed for a total 32GB.
XFX R9 390 DD (no OC)
EVGA Supernova B2 Bronze 750W
Win 7 on an 850 EVO SSD
4 - 7200rpm HDD's


I initially flashed the BIOS on the mobo through USB EZ Flash to get the latest, which was 2013, which was suggested for compatibility with Haswell processors. On build, the system booted to POST with no errors. Mem was initially recognized at 1333, but I set XMP to get the rated 1600. All this went off without a a hitch. I'm just not sure if the array of different types of crashes and freezes I'm getting could be attributed to memory or something else.

I had also installed the r9 390 at the same time as I got the mobo. Due to coil whine I RMA'd the GPU, and the new one installed fine and has no coil whine, so I don't suspect it's related to the problem. It's running on the latest Crimson drivers.

Anyone have an opinion on whether the mem could cause this, or other troubleshooting suggestions?
Thanks

 

[Arbiter1170]

Hey.

So you're able to get the computer up without it instantly BSOD right?

Is there anything you're doing, or any correlation between what programs your running, or how the computer is running, that you've noticed happen before the BSOD?

I know you mentioned memory usage, and sometimes you said it's just random. Is there anything else?
BAD_SPOOL_HEADER, from what I've experienced, can be caused by a whole bunch of different things.

Do you have any programs such a malware bytes? You'd be surprised what can cause BSOD's now a days.

Also, check your ram. Make sure it's right and in the slot correctly.

 

[cin19]

Here is info for "REFERENCE_BY_POINTER" https://msdn.microsoft.com/en-us/library/windows/hardware/ff557386%28v=vs.85%29.aspx it is caused by the driver.
And BAD_SPOOL_HEADER https://msdn.microsoft.com/en-us/library/windows/hardware/ff557389%28v=vs.85%29.aspx

So recommend to update the chipset driver https://downloadcenter.intel.com/download/25094/Chipset-Intel-Chipset-Device-Software-for-Intel-NUC the network driver https://downloadcenter.intel.com/product/70831/Intel-Ethernet-Connection-I217-V also disable the intel onboard iGPU too.

Also try use only one kit the RAM to see the PC run fine or not, when you mixed two kits of RAM, they may or may not cause the problem, that is why we always recommend to buy the kit, only you have to mix them. Also may try manually set the RAM in 9-9-9-27 (low speed) at 1600mhz in BIOS to see the PC runs fine or not. Even set the RAM voltage to 1.55V too. If the PC runs fine back down the voltage to 1.5 or 1.51, 1.52V etc.

Also check your samsung SSD S.M.A.R.T. by the software.

 

[Teemsan]

Here is info for "REFERENCE_BY_POINTER" https://msdn.microsoft.com/en-us/library/windows/hardware/ff557386%28v=vs.85%29.aspx it is caused by the driver.
And BAD_SPOOL_HEADER https://msdn.microsoft.com/en-us/library/windows/hardware/ff557389%28v=vs.85%29.aspx

So recommend to update the chipset driver https://downloadcenter.intel.com/download/25094/Chipset-Intel-Chipset-Device-Software-for-Intel-NUC the network driver https://downloadcenter.intel.com/product/70831/Intel-Ethernet-Connection-I217-V also disable the intel onboard iGPU too.

Also try use only one kit the RAM to see the PC run fine or not, when you mixed two kits of RAM, they may or may not cause the problem, that is why we always recommend to buy the kit, only you have to mix them. Also may try manually set the RAM in 9-9-9-27 (low speed) at 1600mhz in BIOS to see the PC runs fine or not. Even set the RAM voltage to 1.55V too. If the PC runs fine back down the voltage to 1.5 or 1.51, 1.52V etc.

Also check your samsung SSD S.M.A.R.T. by the software.

Thanks cin19

I went to the Intel links you provided and downloaded the updates. I'm not sure if the Chipset driver was the one I already had, because I had updated just a couple months ago. But I went through the install anyways. I did the same with the Network Driver.

I'm kind of confused with the Chipset driver / software install. If I go to the ASUS website support for the Z87 Plus board there are several items under Chipset Drivers, with different names. Here is a screenshot:

So I'm not sure if I additionally need any of those. There is only one item that is listed as Chipset Driver at the bottom. The version number 9.4.0.1017 appears to be older than the one I just got from the Intel site. So I'm guessing that downloading that would be redundant and going backwards??

With the SSD I know SMART, NCQ and TRIM are enabled

I'll try running only one 16 GB (2x8) kit and see what happens. As far as voltage goes, the RAM is rated for 1.5V and I have it manually set to 1.5. So you're saying I should increase the voltage to 1.55 and work down? I'm no expert on RAM but wouldn't undervolting be the way to go to check for stability?

 

[cin19]

You can get the chipset driver from either intel or asus, usually intel has the newer version. I forget you should get the intel ME driver too.

And the SSD is good.

For the RAM, because you said the PC can recognize the RAM at 1333mhz, but has [problem boot after you use the XMP ( 1600mhz), that means the PC may not run stable, if you increase the voltage little bit, that will help, also set the timing to 9-9-9-27 will help too. Try it out. But try only one kit ( 2X 8GB) first. If the PC can run stable then try add other set in. The RAM problem may cause the BAD_SPOOL_HEADER, also the hard drive too that why I ask you check the SSD.

 

[johnbl]

REFERENCE_BY_POINTER, basically every time a driver allocates a object a counter is increased, when it frees a object the counter is decreased.
Screw it up and the counter becomes very large or less than 0 and you can generate a bugcheck.

freeing up objects that you already have freed up can cause the other bugcheck.

I would think the memory dump would name the failing driver name.
you can copy the .dmp files from c:\windows\minidump to a server like microsoft onedrive, share the files for guest access and post a link.

 

[Teemsan]

You can get the chipset driver from either intel or asus, usually intel has the newer version. I forget you should get the intel ME driver too.

And the SSD is good.

For the RAM, because you said the PC can recognize the RAM at 1333mhz, but has [problem boot after you use the XMP ( 1600mhz), that means the PC may not run stable, if you increase the voltage little bit, that will help, also set the timing to 9-9-9-27 will help too. Try it out. But try only one kit ( 2X 8GB) first. If the PC can run stable then try add other set in. The RAM problem may cause the BAD_SPOOL_HEADER, also the hard drive too that why I ask you check the SSD.

Yesterday, just as a test, I reset the BIOS to defaults and just ran the RAM at the default settings as they were recognized by the board - that is 1333 @ 1.5v and the recognized timings of 9-9-9-24. Within about an hour I got a system BSOD with a MEMORY_MANAGEMENT error and a very audible buzzing noise. The system locked up. So I immediately shut it down.

I rebooted and manually set the RAM (one 2x8 - 16GB kit) to your recommended settings of 1.55v and 9-9-9-27 timing @ 1600. So far so good. It's run since yesterday without a hitch. I ran a Prime 95 Blend test for an hour and a half just to test, & it passed with no errors. I'll only be satisfied if this hold s for longer than a week or two without freeze up or BSOD. If these settings hold, should I still then try backing down to 1.4 ... 1.3 etc just to get the voltage as low as possible & still stable? Or is it safe to run 1.5v rated 1600 RAM continuously @ 1.55?

 

[cin19]

1.55V is fine, but you can try back to the 1.5V with 9-9-9-27, even try between 1.50V and 1.55V. But not 1.4V or 1.3V, because lower voltage will cause unstable.

 

[Teemsan]

1.55V is fine, but you can try back to the 1.5V with 9-9-9-27, even try between 1.50V and 1.55V. But not 1.4V or 1.3V, because lower voltage will cause unstable.

Oops, my bad. I meant 1.54 ... 1.53, etc, as you had mentioned those voltages earlier. But now I get the gist of what to do from your last post. I'll report back after a bit when I've run this for a while.

 

[Teemsan]

1.55V is fine, but you can try back to the 1.5V with 9-9-9-27, even try between 1.50V and 1.55V. But not 1.4V or 1.3V, because lower voltage will cause unstable.

Well just when I was thinking I should answer back and say I was in the clear when I had a crash again. It was good for a week, including editing and pushing near the full amount of RAM. doing benches, encodes, multitasking, etc. Then today I was encoding with handbrake. About halfway through the encode I got a Handbrake crash with a hardware CLI.exe error. I restarted Handbrake and tried the encode again.

I was running Handbrake in the background while watching a Youtube vid ... email open, a few browser windows, not much more than about 6 - 8GB's RAM of my 32 GB. The processor was of course running between 80 - 100% with the encode going, but nothing more taxing than I've done many times before. This 4790 can normally multitask like a beast. So then I got a dreaded MEMORY_MANAGEMENT BSOD.

I'd just been running with the RAM at the settings you suggested of 9-9-9-27 timings @ 1.5v. So like I say, I was stable on that for about a week. I added the second 2x8 GB kit yesterday. But maybe I was just hopeful and deceiving myself as the previous BSOD's had subsided to about one every week to two weeks.

Not sure where to go from here. I could just pull the second 2x8 GB kit and run at 16GB's. Not at all ideal for editing as I tend to run at about 24 + GB"S when editing, because I like to have that RAM cache overhead in Vegas for quick Dynamic RAM previews (rather than constantly editing out for quick checks). I was able to achieve this error free for over a year on my old mobo. I may go back to it. I had moved to this mobo to get the extra 2 internal SATA ports. Barring that I could buy a 32GB kit that is actually on the QVL for the board. But for that coin in Canada, or slightly more, I could probably purchase another 8 SATA port mobo where my current RAM is already on the QVL. decisions, decisions ....

 

[cin19]

When you mix two of memory kits, you may or may not have problem like this, unfortunately it does not work for you. May try use the 1.55V for all RAM in1600mhz, or use the RAM in 1333mhz with 9-9-9-27 that means under clock the RAM, if the new RAM is not the option.

 

[Teemsan]

When you mix two of memory kits, you may or may not have problem like this, unfortunately it does not work for you. May try use the 1.55V for all RAM in1600mhz, or use the RAM in 1333mhz with 9-9-9-27 that means under clock the RAM, if the new RAM is not the option.

So far so good after one week running at your above settings:
- Auto detect frequency in BIOS defaulting to an underclock of 1333
- 9-9-9-27 timings
- 1.55 volts

I'll give it another week before I call it totally stable. Any more BSOD's of any kind and I'm not diagnosing it further. I'll buy new RAM. I need a stable platform for productivity and I won't waste more time on it, so fingers crossed this holds.

 

[Teemsan]

When you mix two of memory kits, you may or may not have problem like this, unfortunately it does not work for you. May try use the 1.55V for all RAM in1600mhz, or use the RAM in 1333mhz with 9-9-9-27 that means under clock the RAM, if the new RAM is not the option.

Two weeks stable with these settings:
- 1333mhz
- 9-9-9-27 timings
- 1.55 volts

And that is running CPU and memory intensive editing and encodes, some gaming and running multiple programs. So I'm thinking you nailed it. Thanks for all your help with this cin19.. Much appreciated.