Adding continues
OK. I did a 350 day run and the results are the same. I have to think that the Eat bug was having a significant effect. The problem with going to long double is that the code is much slower. I’m having a play with the Intel complier which will make it much faster but for some reason i keep getting very different netcdf output files. I think these bugs are because i have still got some crap values in the netcdf files.
porosity should be written out as double not long double.
Having another check of the output when double is used instead of long double. As i have fixed a number of bugs i thought it would be good to check the results just in case these bugs were causing the issues. It would be nice to be able o use double as the code runs much faster when doubles are used. I’m not getting any differences. This is great.
Ok. I need to look into the order that things are added in. hmm. I have tried this using the invert types and i get a difference. If i order them in exactly the same order i get the same values.
Have removed the spPREYid = (int *)i_alloc1d(nprey); array.
Ok. Going to do a long run 450 days. If this is ok i will commit the code and then move back to the pwn and jpwn issues.
OK. no inv_id values are used in the ecology lib. Nice.
Getting a core dump at 365 days still. Checking the netcdf files are correct.
OK thats sorted now. I want to clean up the functional group data structures. These have got very ugly.
I want to change the invert type of DC from LAB_DET to CARRION.
Also want to change prey_Avail to no longer require that DLsed, MBsed and DR sed are passed.