See general instructions for tracking down a fault
In Windows 2003 and 2008 server, (and .NET server) all video card acceleration
is turned off by default. In BurnInTest you'll see the error,
3D Graphics test : Error initializing Direct-X device
You need to manually turn it back on if you plan to do video testing. It seems that Microsoft's position is that this acceleration isn't required for typical server functionality. Anyway to fix this,
1. Start the Display applet in Control Panel (go to Start, Control Panel, then click Display).
2. From the General tab, click Advanced, then select the Troubleshoot tab.
3. Move the "Hardware acceleration" pointer to Full.
BurnInTest crashes at the start of testing. This might result in a hardware reset, an invalid page fault or a partial crash with a window being empty (black).
If BurnInTest creates a detailed (minidump) debug file, please email it to us, e.g. C:\Users\<username>\Documents\PassMark\BurnInTest\BurnInTestx64-V8.0 Pro-0016 Beta-20140516-153342-6252-5320.dmp
The most common reason for a fault like this is a problem with the 3D test which uses DirectX 9.0c.
There are a number of video card drivers that don't support hardware accelerated DirectX correctly or have bugs under particular versions of Windows. This can result in a system crash and sometime strange rubbish appearing on the screen. What you should do is:
- Prove that the crash was in fact caused by DirectX. This can be done by turning off the 3D Graphics test in BurnInTest and then checking that this prevents the crash.
- Check that you are using the latest certified video driver for your video card (ask your computer / video card supplier or check their web page).
- Use the "DxDiag" program to verify your drivers and test Direct3D. You should find "DxDiag" in c:\windows\system or in c:\program files\directx\setup.
A number of hardware manufacturers produce video cards which have very poor software support for 3D graphics (DirectX). This is especially the case for the cheaper, bottom of the range video cards. Poor drivers can result in strange visual artifacts appearing on the screen, poor 3D performance and system crashes. The old saying of, "you get what you pay for", seems especially true for video cards, Beware.
Try newer driver versions (and if this doesn't work, try older ones).
We have found that some USB drives can hang (block) on requesting temperature (SMART) information repeatedly. In particular, we have observed this with the Seagate "Expansion External" USB drive. When this occurs the BurnInTest temperature graph will show no results for the USB drive temperature from the point in time that this problem occurs. If you have this problem, it is suggested that you de-select temperature monitoring for this USB drive in BurnInTest Preferences->Temp / Battery.
BurnInTest crashes (with rubbish on the screen or a black screen) when the monitor powersave feature turns on after X minutes of running.
Some video cards' device drivers seem to have a bug that causes this crash when the monitor goes into powersave mode. If you disable the monitor powersave feature (from the Power Management window in the Windows Control panel) this resolves the problem. The main offender seems to be the Maxtrox G400 video card. We have had only 1 report of this problem so it doesn't seem to be very common.
If you are testing serial ports at a high speed, together with USB ports, it should be noted that some chip sets have been reported to have limitations while under load high load. Contact your manufacturer regarding these types of limitations.
This problem can occur when you are running BurnInTest Professional from a networked drive. In this case Windows may fail to load the device driver required to provide access to the parallel port. The only solution, for the moment, is to copy the BurnInTest files to a local drive before running it.
Update1: This has also been seen, even when BurnInTest is on a local drive. It seems to be a timing issue. BurnInTest will automatically attempt to reload the driver after a delay of a few seconds and the test will continue correctly thereafter.
Update2: We have one report that with Gigabyte motherboards you need the latest VIA chip set drivers to get the parallel port test working.
For the parallel port test, the detailed log file indicates some errors, which are not, indicated in the main window of the BurnInTest. Why ?
If a byte is read from the parallel port and it is determined to be wrong, then there is a 2nd attempt to read the same byte (at least 10ms after the 1st attempt). If the 2nd attempt gets the right value then there is no error reported in the main window. If the 2nd attempt also gets a wrong value then the error count in the main window is incremented. In the case of corrupted data being received during the parallel port test the following information is included in the log file.
Sent = the byte value sent to the parallel loopback plug
Rec = the byte value returned from the parallel loopback plug
Att = Read Attempt 1 or 2 (see above)
RawByte = When a byte is read from the pins on the parallel port the bits are in the wrong order. This is the RawByte. After the bits are put back in their correct order then they should match the Sent byte. This is displayed mainly for our debugging purposes.
You need to have administrator privileges to run this test. In Windows 2000, there is no way to disable this security check. Access to Raw Sockets is granted on a per-transport basis. For the address family AF_INET, only administrators have the access necessary to create Raw Sockets.
When running BurnInTest I get a bluescreen crash with the error message DRIVER_IRQL_NOT_LESS_OR_EQUAL in iaStorA.sys.
This appears to be a bug in the Intel RST device driver (iaStorA.sys) that is provoked while trying to read SMART data from a hard disk
(in particular the SMART temperature information) while the disk is under load.
The current workaround is to increase the "sample every x seconds" (on the Temp/Battery tab of the BurnInTest Preference) value to 10 seconds or greater to reduce the chance of this bluescreen occurring or disable the temperature monitoring for the disk.
There is currently a known issue when there is an Intel MIPI camera on a system and both BurnInTest and the Microsoft Windows Store Camera App are launched at the same time.
In same cases BurnInTest is accessing the camera at the same time as the Microsoft camera app and as this access is exclusive the camera app will not be able to access the camera and will not load correctly. The camera is accessed using the "Microsoft Media Foundation" functions, which the Intel driver also uses, currently these functions do not allow shared access of the camera device.
Allowing BurnInTest to finish loading before launching a camera application should prevent this issue from happening.
The most common cause of this error is if a KVM switch is in use or the monitor is turnerd off / goes to sleep. In order for the test to run a monitor must be present on the system.