## Re: [apbs-users] Focusing protocol

 Re: [apbs-users] Focusing protocol From: Baker, Nathan - 2013-04-17 13:03:57 ```Hello -- Your (2c) option is correct: focusing uses a subsequent grids with small lengths but higher resolutions. Thanks, -- Nathan Baker Pacific Northwest National Laboratory Phone: +1-509-375-3997 http://nabaker.me From: Jesper Sørensen > Date: Monday, April 15, 2013 1:13 PM To: APBS-USERS mailing list > Subject: [apbs-users] Focusing protocol Hi all, I am looking at doing some manual focusing calculations and I was wondering what the mg-auto keyword does as a comparison. There are two ways to do it I guess and I just want to make sure which one to use. For the examples sake I'll just use a 2 step focusing calc. Step 1) Start with a large grid (solute fill percentage is medium to low) at resolution X (e.g. 0.5 ang/grid). Then take this grid and supply it to the next calculation which can be, Step 2a) a smaller grid (solute fill is medium to high) at the same resolution X. Step 2b) the same size grid in terms of dimensions x,y,z in angstroms, but with a higher grid resolution If I understand correctly, using 2a, the focusing only improves the effect of the grid edge boundary condition for the smaller grid, avoiding artifacts, but then I've already calculated the solute at a reasonable grid resolution, so why repeat that. Using 2b, will result in a higher grid resolution around the solute, but then I still have to calculate the higher grid resolution for the entire box, which seems costly and ineffective. Might has well just done one calculation with the higher grid resolution, unless the input grid map helps speed up the calculation due to better initial estimates. I guess now that I am writing this, focusing would work really well with a combination of the two. Step 2c) a smaller grid (solute fill is medium to high) and also with a higher grid resolution. Is the latter option 2c actually the one I should use and what does mg-auto do? Best regards, Jesper ```

 [apbs-users] Focusing protocol From: Jesper Soerensen - 2013-04-16 02:03:43 Attachments: text/plain
 Re: [apbs-users] Focusing protocol From: Baker, Nathan - 2013-04-17 13:03:57 ```Hello -- Your (2c) option is correct: focusing uses a subsequent grids with small lengths but higher resolutions. Thanks, -- Nathan Baker Pacific Northwest National Laboratory Phone: +1-509-375-3997 http://nabaker.me From: Jesper Sørensen > Date: Monday, April 15, 2013 1:13 PM To: APBS-USERS mailing list > Subject: [apbs-users] Focusing protocol Hi all, I am looking at doing some manual focusing calculations and I was wondering what the mg-auto keyword does as a comparison. There are two ways to do it I guess and I just want to make sure which one to use. For the examples sake I'll just use a 2 step focusing calc. Step 1) Start with a large grid (solute fill percentage is medium to low) at resolution X (e.g. 0.5 ang/grid). Then take this grid and supply it to the next calculation which can be, Step 2a) a smaller grid (solute fill is medium to high) at the same resolution X. Step 2b) the same size grid in terms of dimensions x,y,z in angstroms, but with a higher grid resolution If I understand correctly, using 2a, the focusing only improves the effect of the grid edge boundary condition for the smaller grid, avoiding artifacts, but then I've already calculated the solute at a reasonable grid resolution, so why repeat that. Using 2b, will result in a higher grid resolution around the solute, but then I still have to calculate the higher grid resolution for the entire box, which seems costly and ineffective. Might has well just done one calculation with the higher grid resolution, unless the input grid map helps speed up the calculation due to better initial estimates. I guess now that I am writing this, focusing would work really well with a combination of the two. Step 2c) a smaller grid (solute fill is medium to high) and also with a higher grid resolution. Is the latter option 2c actually the one I should use and what does mg-auto do? Best regards, Jesper ```
 Re: [apbs-users] Focusing protocol From: Jesper Sørensen - 2013-04-17 17:29:59 ```Hi Nathan, Is there a simple formula/protocol that mg-auto uses? Or is that a very complex system dependent protocol? Best, Jesper On Apr 17, 2013, at 6:03 AM, "Baker, Nathan" wrote: > Hello -- > > Your (2c) option is correct: focusing uses a subsequent grids with small lengths but higher resolutions. > > Thanks, > > -- > Nathan Baker > Pacific Northwest National Laboratory > Phone: +1-509-375-3997 > http://nabaker.me > > From: Jesper Sørensen > > Date: Monday, April 15, 2013 1:13 PM > To: APBS-USERS mailing list > > Subject: [apbs-users] Focusing protocol > > Hi all, > > I am looking at doing some manual focusing calculations and I was wondering what the mg-auto keyword does as a comparison. > There are two ways to do it I guess and I just want to make sure which one to use. > For the examples sake I'll just use a 2 step focusing calc. > > Step 1) Start with a large grid (solute fill percentage is medium to low) at resolution X (e.g. 0.5 ang/grid). Then take this grid and supply it to the next calculation which can be, > > Step 2a) a smaller grid (solute fill is medium to high) at the same resolution X. > Step 2b) the same size grid in terms of dimensions x,y,z in angstroms, but with a higher grid resolution > > If I understand correctly, using 2a, the focusing only improves the effect of the grid edge boundary condition for the smaller grid, avoiding artifacts, but then I've already calculated the solute at a reasonable grid resolution, so why repeat that. > Using 2b, will result in a higher grid resolution around the solute, but then I still have to calculate the higher grid resolution for the entire box, which seems costly and ineffective. Might has well just done one calculation with the higher grid resolution, unless the input grid map helps speed up the calculation due to better initial estimates. > > I guess now that I am writing this, focusing would work really well with a combination of the two. > Step 2c) a smaller grid (solute fill is medium to high) and also with a higher grid resolution. > > Is the latter option 2c actually the one I should use and what does mg-auto do? > > Best regards, > Jesper > ```
 Re: [apbs-users] Focusing protocol From: Baker, Nathan - 2013-04-18 03:24:38 ```Hi Jesper -- It's not too complicated. Take a look at the setupCalcMGAUTO() code in APBS at http://sourceforge.net/p/apbs/code/ci/master/tree/src/generic/nosh.c#l1552 -- this outlines the general steps for breaking an mg-auto calculation down into individual mg-manual parts. Let me know if you have any questions. Thanks, -- Nathan Baker Pacific Northwest National Laboratory Phone: +1-509-375-3997 http://nabaker.me On 4/17/13 10:29 AM, "Jesper Sørensen" wrote: >Hi Nathan, > >Is there a simple formula/protocol that mg-auto uses? Or is that a very >complex system dependent protocol? > >Best, >Jesper > > >On Apr 17, 2013, at 6:03 AM, "Baker, Nathan" >wrote: > >> Hello -- >> >> Your (2c) option is correct: focusing uses a subsequent grids with >>small lengths but higher resolutions. >> >> Thanks, >> >> -- >> Nathan Baker >> Pacific Northwest National Laboratory >> Phone: +1-509-375-3997 >> http://nabaker.me >> >> From: Jesper Sørensen > >> Date: Monday, April 15, 2013 1:13 PM >> To: APBS-USERS mailing list >>> >> Subject: [apbs-users] Focusing protocol >> >> Hi all, >> >> I am looking at doing some manual focusing calculations and I was >>wondering what the mg-auto keyword does as a comparison. >> There are two ways to do it I guess and I just want to make sure which >>one to use. >> For the examples sake I'll just use a 2 step focusing calc. >> >> Step 1) Start with a large grid (solute fill percentage is medium to >>low) at resolution X (e.g. 0.5 ang/grid). Then take this grid and supply >>it to the next calculation which can be, >> >> Step 2a) a smaller grid (solute fill is medium to high) at the same >>resolution X. >> Step 2b) the same size grid in terms of dimensions x,y,z in angstroms, >>but with a higher grid resolution >> >> If I understand correctly, using 2a, the focusing only improves the >>effect of the grid edge boundary condition for the smaller grid, >>avoiding artifacts, but then I've already calculated the solute at a >>reasonable grid resolution, so why repeat that. >> Using 2b, will result in a higher grid resolution around the solute, >>but then I still have to calculate the higher grid resolution for the >>entire box, which seems costly and ineffective. Might has well just done >>one calculation with the higher grid resolution, unless the input grid >>map helps speed up the calculation due to better initial estimates. >> >> I guess now that I am writing this, focusing would work really well >>with a combination of the two. >> Step 2c) a smaller grid (solute fill is medium to high) and also with a >>higher grid resolution. >> >> Is the latter option 2c actually the one I should use and what does >>mg-auto do? >> >> Best regards, >> Jesper >> > >-- >You received this message because you are subscribed to the Google Groups >"APBS users" group. >To unsubscribe from this group and stop receiving emails from it, send an >email to apbs-users+unsubscribe@... >For more options, visit https://groups.google.com/groups/opt_out. > > ```