dynamic exception specifications are deprecated in C++11; use ‘noexcept’ instead [-Wdeprecated]
here.
on an unreleated but IMO interesting note, GCC 7 seems to become a great release for our purposes: on my machine the linearization performance improved by about 20% using flow_ebos with the norne deck when compared to GCC 5.4.0:
================ End of simulation ===============
Total time taken (seconds): 661.888
Solver time (seconds): 658.491
Assembly time (seconds): 284.158
Linear solve time (seconds): 313.216
Update time (seconds): 9.94308
Output write time (seconds): 0.00081154
Overall Well Iterations: 890
Overall Linearizations: 1997
Overall Newton Iterations: 1653
Overall Linear Iterations: 24167
these figures need to taken with a grain of salt, though: I had to disable output because it segfaulted (this is probably related to OPM/opm-simulators#1046) and I had to recompile boost because of linker errors. the latter caused me to loose another 3 hours (or so).
GCC 7 warns
here.
on an unreleated but IMO interesting note, GCC 7 seems to become a great release for our purposes: on my machine the linearization performance improved by about 20% using flow_ebos with the norne deck when compared to GCC 5.4.0:
these figures need to taken with a grain of salt, though: I had to disable output because it segfaulted (this is probably related to OPM/opm-simulators#1046) and I had to recompile boost because of linker errors. the latter caused me to loose another 3 hours (or so).