Announcement

**David (PassMark)** · Dec-18-2023, 08:55 AM

Relevant part of the log is,

SERIOUS: 2023-12-17 00:06:00, 3D Graphics, An error occured during the DX12 3D test
LOG NOTE: 2023-12-17 00:06:00, 3D Graphics, Unexpected error running DirectX 12 Test. Error Number 0x887A0005 (-2005270523)
LOG NOTE: 2023-12-17 00:06:00, 3D Graphics, The GPU device instance has been suspended. Use GetDeviceRemovedReason to determine the appropriate action.
LOG NOTE: 2023-12-17 00:06:00, 3D Graphics, DXGI_ERROR_DEVICE_REMOVED (0x887a0005) - Device Removed reason: DXGI_ERROR_DEVICE_HUNG (0x887a0006)
SERIOUS: 2023-12-17 00:06:00, 2D Graphics, Error writing to GPU memory
LOG NOTE: 2023-12-17 00:06:00, 2D Graphics, GPU Intel(R) Arc(TM) Graphics, Failed to write GPU buffer object (-5) - Block 73 / 100

And the errors in the user interface was.
GPGPU 993 196 Trillion FAIL 1 No operations reported in timeout period
2D Graphics 5922 5.405 Trillion FAIL 1 Error writing to GPU memory
3D Graphics 36 1.104 Million FAIL 1 An error occured during the DX12 3D test

This happened at around 8 hours into the test.

As noted in the log, from the applications point of view the entire video card disappeared (was removed from the system). The sub-reason, was a hang. The event log also reports a crash with the video card device driver.

I assume if you repeat the test, the same thing happens again, at around the 8 hour mark?
If this was the case then it might be a resource leak in the video card driver eventually causing a hang.

There might be a few more details available if you turn on level 2 logging in BurnInTest. The log file will be much much bigger however.

Might also be interesting to repeat the same test without the Memory(RAM) test being active, to see if the issue was related to low RAM.

**Stanley Shen** · Dec-20-2023, 12:54 AM

I test it again and without RAM test. test result still 2D/3D/GPGPU fail. I will use Level 2 logging to test again and provide the log later

Attached Files

1220.zip (366.6 KB, 231 views)

**David (PassMark)** · Dec-20-2023, 01:19 AM

New log showed almost the same thing. A crash in the video card driver. The sub-reason was different however. This time, it was device reset.
Also the crash was 13 hours into the test not 8.

LOG NOTE: 2023-12-20 04:30:29, 3D Graphics, DXGI_ERROR_DEVICE_REMOVED (0x887a0005) - Device Removed reason: DXGI_ERROR_DEVICE_RESET (0x887a0007)

Next step would be checking the level 3 log and looking for excess RAM usage.

**Stanley Shen** · Dec-22-2023, 07:01 AM

Provide Level 2 log, because passmark limitation, i pull data on google Drive. please download the log on below linking.
https://drive.google.com/file/d/13Ku...usp=drive_link
please help us to anlysis this 2D/3D/GPGPU fail on Intel MTL-H platform

**Stanley Shen** · Dec-28-2023, 02:01 AM

i delete some pass log to downgrade size, please help us analysis level 2 log to find 2D/3D/GPGPU fail root cause

Attached Files

1222.zip (457.5 KB, 112 views)

**David (PassMark)** · Dec-28-2023, 03:59 AM

Doesn't look like the machine ran out of physical RAM (via a leak), which I thought might be a possibility.

So as far was we can tell the video card driver crashed or locked up.
LOG NOTE: 2023-12-22 08:27:58, 3D Graphics, DXGI_ERROR_DEVICE_REMOVED (0x887a0005) - Device Removed reason: DXGI_ERROR_DEVICE_HUNG (0x887a0006)

Shortly after this BurnInTest started reporting errors, as you would expect if the video card suddenly disappeared.

We don't know why the video card driver crashed. A bug of some sort we presume.

Announcement

MTL-H 2D 3D GPGPU fail

MTL-H 2D 3D GPGPU fail

Comment

Comment

Comment

Comment

Comment

Comment