grid list: Too many quadrature shells
Just Got Here
7:41:24 AM PDT - Mon, Jul 2nd 2012
Good morning,

I'm trying to run a large DFT calculation and I get the following error message:

grid_list: Too many quadrature shells.


followed by:

grid_list: raise MXqshells to     90001


Any idea how I can fix this? (I can't find anything about MXqshells in the manual).

Thanks,

Vincent

Forum Vet
2:29:33 PM PDT - Mon, Jul 2nd 2012
Hi Vincent,

This must be a very large system, you are the first running into this limit as far as I know. How many atoms/basis functions is the system you are trying to do?

There is no simple environment variable that can be set. You will have to go to src/nwdft/include/cdft.h and increase some parameters. In this file the MXqshells is around line 106. After changing this you should in src/nwdft do a "make clean" followed by the normal "make ..." command you use. Then in src/ do a "make link".

Bert


Quote:Vmeunier Jul 2nd 2:41 pm
Good morning,

I'm trying to run a large DFT calculation and I get the following error message:

grid_list: Too many quadrature shells.


followed by:

grid_list: raise MXqshells to     90001


Any idea how I can fix this? (I can't find anything about MXqshells in the manual).

Thanks,

Vincent

Just Got Here
3:19:19 PM PDT - Mon, Jul 2nd 2012
Yes, it is a big big system (a couple thousand atoms). I'm using a baby basis-set (sto-3g) but, frankly, I was simply trying to push the limits of what I could do with NWCHEM.

I'm using a version of NWCHEM that has been installed by the sysadmin and I'm not sure it is easy to fix the header file and recompile.

I think my best option is probably to try to reduce the system size!

Thanks for your quick help!

Vincent


Forum >> NWChem's corner >> Running NWChem