Pyrolysis Simulation Time

[johodges]

This input file results in a significant difference in the computational time between FDS version 6.2 and FDS version 6.7. From the output files it seems the majority of the difference is in the wall routine. The total run time on my machine in 6.2 is 33 seconds versus 410 seconds in 6.7. The input file is simulating ignition through a cone test.

pyroSimTime.zip

[mcgratta]

This should be fixed. Recompile the latest source code and let me know if the problem is fixed.

[johodges]

Thanks for looking into this issue. I compiled the latest version and re-ran the comparison. The wall time is lower but the DIVG, MASS, and VELO times are higher than before. Please see the attached output/cpu files.
comparisonzip.zip

[mcgratta]

Your new times are much higher. Are you sure you compiled the release, as opposed to debug, version. When I run the case, it is slightly slower than the 6.2 times, but probably just because of a difference in processor speed.

[johodges]

Thanks for pointing that out, the increase in other times was a result of compiling the debug version. When I run the comparison with the release version the times are comparable. Thanks for the support.

[mcgratta]

No problem. You caught a nasty bug.