Previously, if the plotting library (which we aren't currently using) was included on a machine that did not have it installed, the test would fail, but the test would still pass because we were comparing our return code with the return code of running the benchmark code.
All return code are not compared with 0 and fail on non zero value.
Commented out references to this library.
Previously, if the plotting library (which we aren't currently using) was included on a machine that did not have it installed, the test would fail, but the test would still pass because we were comparing our return code with the return code of running the benchmark code.
All return code are not compared with 0 and fail on non zero value. Commented out references to this library.