Template Resets Machine Upon Loading

May 28, 2009 at 8:27 PM

Whenever I load the template, it resets my machine.  I see the input fields appear and then after a significant delay I see the "Document Actions" box.  Immediately thereafter the machine does a hard reset (screen goes immediately black -- no error message, no shutdown process).  I have tried uninstalling/reinstalling/repairing NodeXL, the .NET framework, .NET SP1, Office 2007 SP2, Compatibility Pack SP2, etc., etc. as well as disabling/uninstalling other add-ins, all without any success.  Does anyone have any insights into this?

May 28, 2009 at 9:50 PM

This one shocks me.  It's difficult nowadays for application software to cause a hard reset.  On my own software development machines, I've seen this happen exactly twice since Windows XP was released, and both instances were caused by faulty low-level security software.

In other words, I'm hard-pressed to explain how NodeXL could cause this.

A few questions:

1. Was your machine stable until you installed NodeXL and attempted to open its template?

2. Was this the first time you installed NodeXL, or were you upgrading to a new release?

3. Do you see anything unusual in the Windows Event Log?  On Vista, you can get there by typing "Event Viewer" into the Start menu's Search box.  On XP, you can get there from here: http://support.microsoft.com/kb/308427.  You would want to search for an "Error" event in the Application, Security, or System logs that occurred at the time you opened the template.

-- Tony

May 28, 2009 at 9:53 PM

One other question:

4. How much memory does your machine have?  NodeXL is a memory hog with large workbooks, although I wouldn't expect memory problems with the empty template and I wouldn't expect a hard reset in any case.

-- Tony

May 29, 2009 at 2:44 PM

Yes, this is surprising!  In response to the above:

1: The machine was and continues to be rock solid stable in every other way.  Not even so much as an error message.

2: This was a new install.  During the process of trying to resolve this I ended up trying both 1.0.1.73 and .85 and experienced the same problem with both.

4: (Since the answer to 3 is very long) 4GB

3: Yes, I do see messages that look relevant, but they do not appear every time the machine has been reset:

In one instance, from the Application Event Log:

[ERROR] Faulting application excel.exe, version 12.0.6425.1000, stamp 49d64dd6, faulting module mso.dll, version 12.0.6425.1000, stamp 49d65443, debug? 0, fault address 0x001371e0.

Followed by:

[INFORMATION] Bucket 1295037206, bucket table 1, faulting application excel.exe, version 12.0.6425.1000, stamp 49d64dd6, faulting module mso.dll, version 12.0.6425.1000, stamp 49d65443, debug? 0, fault address 0x001371e0.

Then, after reset, in the Sytem Event Log:

[INFORMATION / Source: "Save Dump"] The computer has rebooted from a bugcheck.  The bugcheck was: 0x1000007e (0xc0000005, 0xbfc34800, 0xa9b4150c, 0xa9b41208). A dump was saved in: C:\WINDOWS\Minidump\Mini052809-08.dmp.

Then, mysterious and inconsistent error messages upon reboot:

[INFORMATION / "Application Popup"] Application popup:  : \SystemRoot\System32\RDPDD.dll failed to load

 

 

So, looking in the .DMP file from above I see:

BugCheck 1000007E, {c0000005, bfc34800, a9b4150c, a9b41208}

...

Probably caused by : ati3duag.dll ( ati3duag+106800 )

 

....

 

Debugging Details:
------------------


EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at "0x%08lx" referenced memory at "0x%08lx". The memory could not be "%s".

FAULTING_IP:
ati3duag!SWVS_LoadSrc+150
bfc34800 8b4604          mov     eax,dword ptr [esi+4]

EXCEPTION_RECORD:  a9b4150c -- (.exr 0xffffffffa9b4150c)
ExceptionAddress: bfc34800 (ati3duag!SWVS_LoadSrc+0x00000150)
   ExceptionCode: c0000005 (Access violation)
  ExceptionFlags: 00000000
NumberParameters: 2
   Parameter[0]: 00000000
   Parameter[1]: 0000000a
Attempt to read from address 0000000a

CONTEXT:  a9b41208 -- (.cxr 0xffffffffa9b41208)
eax=00000000 ebx=e509f720 ecx=e4e7c040 edx=e4e53074 esi=00000006 edi=a9b419a4
eip=bfc34800 esp=a9b415d4 ebp=e4e53040 iopl=0         nv up ei pl nz na pe nc
cs=0008  ss=0010  ds=0023  es=0023  fs=0030  gs=0000             efl=00010206
ati3duag!SWVS_LoadSrc+0x150:
bfc34800 8b4604          mov     eax,dword ptr [esi+4] ds:0023:0000000a=????????
Resetting default scope

PROCESS_NAME:  EXCEL.EXE

ERROR_CODE: (NTSTATUS) 0xc0000005 - The instruction at "0x%08lx" referenced memory at "0x%08lx". The memory could not be "%s".

EXCEPTION_PARAMETER1:  00000000

EXCEPTION_PARAMETER2:  0000000a

READ_ADDRESS:  0000000a

FOLLOWUP_IP:
ati3duag!SWVS_LoadSrc+150
bfc34800 8b4604          mov     eax,dword ptr [esi+4]

BUGCHECK_STR:  0x7E

DEFAULT_BUCKET_ID:  NULL_CLASS_PTR_DEREFERENCE

LAST_CONTROL_TRANSFER:  from 00000000 to bfc34800

STACK_TEXT: 
a9b415e0 00000000 e4e7c040 00000007 00000001 ati3duag!SWVS_LoadSrc+0x150


SYMBOL_STACK_INDEX:  0

SYMBOL_NAME:  ati3duag!SWVS_LoadSrc+150

FOLLOWUP_NAME:  MachineOwner

MODULE_NAME: ati3duag

IMAGE_NAME:  ati3duag.dll

DEBUG_FLR_IMAGE_TIMESTAMP:  46eb3954

STACK_COMMAND:  .cxr 0xffffffffa9b41208 ; kb

FAILURE_BUCKET_ID:  0x7E_ati3duag!SWVS_LoadSrc+150

BUCKET_ID:  0x7E_ati3duag!SWVS_LoadSrc+150

Followup: MachineOwner

 

So, ATI3DUAG.DLL is one of my ATI Radeon HD 2400 XT's driver files.  Regular graphs don't cause Excel to crash -- but does NodeXL invoke other video modes that could cause this?

May 29, 2009 at 3:40 PM

Thank you!  I wish all problem reports were as detailed as yours.

Since version 1.0.1.71, NodeXL has used a newer drawing technology called Windows Presentation Foundation, or WPF.  Unlike the older technology used by the built-in Excel graphs, WPF takes advantage of the graphic acceleration capabilities of well-spec'd video cards such as yours.  Given that the fault appears to be occurring in the video card's driver, my first guess is that the driver has a bug and that the bug rears its ugly head when WPF attempts to use one of its advanced features.

So my next question is this:  Do you have ATI's latest driver for the card?  If I'm reading their site correctly, the latest driver is dated 5/15/2009.

-- Tony

May 29, 2009 at 5:01 PM

I updated the driver and it is no longer resetting!  Everything appears to be working fine.

 

Thank you for your help in getting to the bottom of this!

 

 

May 29, 2009 at 5:25 PM

You've made my weekend.  I was pretty worried when I first read your post.  Now I know what to recommend first if this happens to anyone else.

-- Tony