[Bug] autotvm_matmul Check failed:(code == RPCCode::kReturn)is false: code =kShutdown

[MalaJeans]

When using the official tutorial: autotvm_matmul_x86.py, the tuning trials is set to 10,
and the matrix size is set to N, L, M = 204320432043. An error occurred during the execution as follows:
Check failed:(code == RPCCode::kReturn)is false: code =1
Check failed:(code == RPCCode::kReturn)is false: code =kShutdown
How can I solve this problem?

My environment is:
Ubuntu 20.04.6

Thanks!

[Johnson9009]

It seems the RPC server timeout when executing, so try to increase the "session_timeout".

[MalaJeans]

@Johnson9009
Change the session_timeout of rpc.connect(), right？
In which file rpc.connect() needs to be modified?
Thanks!

[Johnson9009]

@Johnson9009 Change the session_timeout of rpc.connect(), right？ In which file rpc.connect() needs to be modified? Thanks!

Yes, just try increase the session_timeout of rpc.connect, if you use RPC tracker, it is the session_timeout of "request".

    tracker = rpc.connect_tracker(tracker_host, tracker_port)
    remote = tracker.request(key, priority=0, session_timeout=60)

[Johnson9009]

Please change the tile of your issue, so others can know what the problem about and help you.

[MalaJeans]

OK, thanks！I will have a try!

[MrGouYou]

@Johnson9009 Change the session_timeout of rpc.connect(), right？ In which file rpc.connect() needs to be modified? Thanks!

Yes, just try increase the session_timeout of rpc.connect, if you use RPC tracker, it is the session_timeout of "request".

    tracker = rpc.connect_tracker(tracker_host, tracker_port)
    remote = tracker.request(key, priority=0, session_timeout=60)

Actually, i am running with 'LocalRunner' (device is cuda, llvm is ok), and also get similar problem.