Error: clEnqueueNDRangeKernel

[FIFER]

Hello,
I was running the align Photo step on a large dataset  (27k pics).
I got this error some 12 h into the pair selection step.

Error: clEnqueueNDRangeKernel(queue(), kernel.kernel(), work_dim, global_work_offset, global_work_size, local_work_size, 0, NULL, &ev): CL_INVALID_GLOBAL_WORK_SIZE

What could be the cause?

Attached the log of the processing

[Alexey Pasumansky]

Hello FIFER,

What version of Metashape you are using and what GPU driver version is installed on the system?

Do you also observe similar problems on smaller datasets, like a couple hundreds photos?

[FIFER]

Hello Alexey,

It was the first time I got that error. So far worked ok with datasets up to 13k images.

Here the details you asked:

Metashape is 1.8.4
the GPU is a Radeon Instinct MI25 MxGPU
driver version is 27.20.21021.5007 (Advanced Micro Devices, Inc.

GPU 0

   Radeon Instinct MI25 MxGPU

   Driver version:   27.20.21021.5007
   Driver date:   12/1/2021
   DirectX version:   12 (FL 12.1)
   Physical location:   Virtual PCI Bus Slot 0 Serial 495617389


   Dedicated GPU memory   0.1/16.0 GB
   Shared GPU memory   0.0/56.0 GB
   GPU Memory   0.2/72.0 GB

CPU

   AMD EPYC 7V12 64-Core Processor

   Base speed:   2.44 GHz

   Virtual processors:   32
   Virtual machine:   Yes


   Utilization   4%
   Speed   2.44 GHz