Properly initialize ConvSolver by default.

[CAHEK7]

Based on #2379 (comment)

Currently ConvSolver initialized with miopenStatusSuccess by default, but it's not actually initialized with meaningful values:

    ConvSolution(miopenStatus_t status_ = miopenStatusSuccess)
        : status(status_),
          solver_id("<unknown>"),
          invoker_factory(boost::none),
          workspace_sz(0),
          grp_tile1(-1),
          grp_tile0(-1),
          in_tile1(-1),
          in_tile0(-1),
          out_pix_tile1(-1),
          out_pix_tile0(-1),
          n_out_pix_tiles(-1),
          n_in_data_tiles(-1),
          n_stacks(-1)
    {}

It's better to set the default status to miopenStatusNotInitialized, because otherwise empty solver may be used as a normal one and may trigger errors much later (or the other values will be checked to verify that it's a properly initialized solver, which make the status_ meaningless).

@atamazov what do you think?

[atamazov]

@CAHEK7 Please remove https://github.com/ROCmSoftwarePlatform/MIOpen/labels/IMPACTS_API.

wrt initialization of status

• at first glance your proposal is good
• however AFAICS a Solver should either return a valid Solution or flag an error.
• therefore inclusion of status into Solution seems to have been the wrong choice
• the above is the reason of the bizarre initialization of status
• but it works well

Therefore, if we are not going to change the structure of ConvSolution, then I would keep current initialization of status as is.

[JehandadKhan]

@CAHEK7 If you agree with the comment above, please close the issue.

[CAHEK7]

We talked one more time (with @atamazov and @DrizztDoUrden) and decided to remove the status from ConvSolution entirely and either return correctly initialized structure or throw an exception.
But it's not a critical task right now.

[CAHEK7]

All of the CK, MLIR and GEMM based kernels do return {}; if CK\MLIR\GEMM disabled and must do return {miopenStatusNotInitialized}.

Current design is error-prone and must be fixed.

[ppanchad-amd]

@CAHEK7 Is this issue resolved? Thanks!

[CAHEK7]

@ppanchad-amd no. It is still there.