We encounted gpu hang on
If you are a silicon partner or OEM I would suggest raising a support case with Arm - it's unlikely we'll be able to help much via the forums as this is some internal error in the driver, and unlikely to be an application issue.
If you are an app developer can you provide more details about what you are doing? Is the failure reproducible? On which device? Can you share a reproducer?
hi,
I had the same problem with the same GPU this morning. the problem start after a loop of more than 20 kernel using host prt pointer of (256*47*sizeof(cl_int4).
first error was on enqueuemapbuffer -14 and on the next image processed error -36.
using uchar4 and ushort4 do not give any trouble even with a loop of 35 kernel.
Even with (256*55*sizeof(cl_ushort4) and a loop of 53 kernel everythings work.
i found the same problem using ushort8.