MFXVideoCORE_SyncOperation still fails with MFX_ERR_GPU_HANG

[gonta]

Hi,

I have read driver version 9955 (30.0.100.9955) no longer has MFX_ERR_GPU_HANG issue. 

I use more later version of driver, but still got MFX_ERR_GPU_HANG error after MFXVideoCORE_SyncOperation() hangs about 100 sec.

I decode several streams in parallel.

What I want to know:

1. Is there any workaround for the issue?

2. I know that resetting the decoder will restore it, but I cannot wait 100 seconds for the error to occur. Let me know where do you define the time to detect GPU hangup. I want shorten it.

Environment:

Processor:12th Gen Intel(R) Core(TM) i7-12700 2.10 GHz
RAM: 32.0 GB (31.7 GB available)
Windows 11 Pro  21H2  (22000.348)
Intel(R) UHD Graohics 770: 30.0.101.1002
oneVPL: 2023.0.0

Thank you.

[gonta]

It was useless to set a short timeout period for MFXVideoCORE_SyncOperation().
MFXVideoCORE_SyncOperation() itself now returns with MFX_WRN_IN_EXECUTION at the set timeout period, but it eventually hangs for about 100 seconds when resetting the decoder.

Thank you.

[AthiraM_Intel]

Hi,

Thank you for posting in Intel Communities.

Could you please share the driver version you are using?

Thank you

[gonta]

Hi,

I use 30.0.101.1002

thank you

[AthiraM_Intel]

Hi,

Thank you for the quick response. We are checking on this internally.

Could you please share the minimal reproducer for your application and the exact steps, so that we can reproduce the same from end.

or

Reproduce the same with one of our sample and let us know the details.

Thanks

[AthiraM_Intel]

Hi,

We have not heard back from you. Could you please give us an update?

Thanks

[gonta]

Hi,

I'm sorry but it's not easy to offer the sample to reproduce the issue.

Moreover, the issue is very unreproducible, occur only once every few days.

Is there any way to reduce the time it takes for MFX_ERR_GPU_HANG to return?

thank you

[AthiraM_Intel]

Hi,

We are checking on this internally. We will get back to you with an update.

Thanks

[AthiraM_Intel]

Hi,

GPU hang is not predictable and should not happen in normal situation.

If we know your workload and use case a bit more, then we might be able to help you to avoid. If not, it might be hard to provide necessary information.

It seems like you are using ADL (https://ark.intel.com/content/www/us/en/ark/products/134591/intel-core-i712700-processor-25m-cache-up-to-4-90-ghz.html)

And the driver version is old.

Please use latest driver from  https://www.intel.com/content/www/us/en/download/741626/intel-arc-pro-graphics-windows.html

It might have some critical fixes that causes GPU hang.

If you face any further issue, please let us know.

Thanks

[gonta]

Thank you for response.

I use latest driver (31.0.101.4316), but I still get the issue.

thank you

[AthiraM_Intel]

Hi,

To assist you further we need sample reproducer code. Could you please share sample reproducer code and the exact steps you followed so that we can reproduce the issue from our side.

Thanks

[AthiraM_Intel]

Hi,

We have not heard back form you. Could you please share sample reproducer code and the exact steps you followed so that we can help you further.

Thanks

[AthiraM_Intel]

Hi,

We have not heard back from you. This thread will no longer be monitored by Intel. If you need further assistance, please post a new question.

Thanks