Autoscheduler issue for mobile GPU

Troubleshooting
#1

Hi! while trying to run autoscheduler for mobile GPU, I am facing an error

Time elapsed for measurement: 2.30 s
----------------------------------------------------------------------
------------------------------  [ Done ]
----------------------------------------------------------------------
No valid state found in this search round. Check if it has traversed all of the search space.


Traceback (most recent call last):
  File "/home/user/workspace/tvm-scripts/addmm_auto.py", line 103, in <module>
    sch, args = task.apply_best(log_file)
  File "/home/user/.local/lib/python3.10/site-packages/tvm-0.13.dev183+g4267fbf6a-py3.10-linux-x86_64.egg/tvm/auto_scheduler/search_task.py", line 516, in apply_best
    raise RuntimeError(
RuntimeError: Cannot find any valid schedule for ["matmul_add", 1024, 1024, 1024, "float32"] in file matmul_add.json

Device is available

Create task...
Success
Computational DAG:
A = PLACEHOLDER [1024, 1024]
B = PLACEHOLDER [1024, 1024]
matmul(i, j) += (A[i, k]*B[k, j])
C = PLACEHOLDER [1024, 1024]
out(i, j) = (matmul[i, j] + C[i, j])

Begin tuning...
Get devices for measurement successfully!
.T.TTime elapsed for measurement: 18.66 s

Target info

target = tvm.target.Target("opencl -device=mali", host="llvm -mtriple=aarch64-unknown-linux-android33")

Can you please help with this issue?