On Mon, Jul 23, 2018 at 1:46 PM, Christoph Junghans <[email protected]> wrote:
> On Mon, Jul 23, 2018 at 6:39 AM, 'Andrey Brukhno' via votca > <[email protected]> wrote: > > On Mon, Jul 23, 2018 at 1:17 PM, Christoph Junghans <[email protected]> > > wrote: > >> > >> On Mon, Jul 23, 2018 at 2:39 AM, 'Andrey Brukhno' via votca > >> <[email protected]> wrote: > >> > > >> > On Mon, Jul 23, 2018 at 2:22 AM, Christoph Junghans < > [email protected]> > >> > wrote: > >> >> > >> >> On Sun, Jul 22, 2018 at 1:27 PM, 'Andrey Brukhno' via votca > >> >> <[email protected]> wrote: > >> >> > Thanks for your reply Christoph. > >> >> > > >> >> > I did check it, but inverse.log does not contain sufficient info, > see > >> >> > the > >> >> > extract below. > >> >> > I will be investigating it further to make sure I did not do > anything > >> >> > stupid > >> >> > myself, although I haven't changed much for the test iteration > except > >> >> > recompiling the tpr files for the newer version of Gromacs). > >> >> The " rdf calculation is already done" tells me that this is not a > run > >> >> from scratch. > >> >> Try a fresh run, because some chance in the xml don't get populated > >> >> until the next iteration step. > >> >> And grep for "csg_stat" (not rdf), you should see actually command, > >> >> which got run, e.g.: > >> >> "Running critical command 'csg_stat --nt 4 --options settings.xml > >> >> --top topol.tpr --trj traj.xtc --begin 20 --first-frame 0' " > >> >> From that you can see if all the options are read as excepted. > >> >> > >> > > >> > I knew you were going to say this... > >> > As a matter of fact, there is no such message from > calc_rdf_generic.sh, > >> > but > >> > instead: > >> > > >> > msg "Calculating rdfs with csg_stat using $tasks tasks" > >> > critical csg_stat --nt $tasks --options "$CSGXMLFILE" --top "$topol" > >> > --trj > >> > "$traj" --begin $equi_time --first-frame $first_frame ${error_opts} > >> > ${maps:+--cg ${maps}} > >> > mark_done "rdf_calculation${suffix}" > >> Well the "critical" function includes the print statement: > >> critical() { > >> [[ $quiet = "no" ]] && echo "Running critical command '$*'" >&2 > >> "$@" || die "${FUNCNAME[0]}: '$*' failed" > >> } > >> So I am not sure when that got lost in your case. > >> > > >> > so I had to add this after the msg and before the csg_stat call: > >> > echo "Doing: critical csg_stat --nt $tasks --options "$CSGXMLFILE" > >> > --top > >> > "$topol" --trj "$traj" --begin $equi_time --first-frame $first_frame > >> > ${error_opts} ${maps:+--cg ${maps}}" > >> > > >> > This shows me that topol-rdf.tpr is actually used (extract from > >> > inverse.log): > >> > > >> > Make update for ibi > >> > Calculating rdfs with csg_stat using 4 tasks > >> > Doing: critical csg_stat --nt 4 --options > >> > > >> > /home/andrey/Work/Models/DOPC-long-wtip4p-new/ibi-Rc1.65nm- > freeN-dihs_CG-fits1-151cT/dopc_cg-int-map11.xml > >> > --top topol-rdf.tpr --trj traj.xtc --begin 0 --first-frame 0 > >> > --block-length > >> > 1000000 --ext dist.block --cg > >> > > >> > /home/andrey/Work/Models/DOPC-long-wtip4p-new/ibi-Rc1.65nm- > freeN-dihs_CG-fits1-151cT/dopc_cg-map11.xml; > >> > begin to calculate distribution functions > >> > > >> > However, the distributions are still wrong, and I am lost now as to > >> > which > >> > program to blame (again all the initial files are the same as in my > >> > previous > >> > iteration). > >> > I compared the outputs of `gmx dump topol-rdf.tpr' and `gmx dump > >> > topol.tpr' > >> > generated with the new Gromacs and everything seems alright, the > >> > relevant > >> > differences in exclusion lists are there. > >> > The problem is also that the original calculation was done on a > cluster > >> > more > >> > than a year ago, so I cannot reproduce exactly all the installation > >> > setup, > >> > not on my desktop anyway. But I know that those calculations were > >> > correct > >> > and well converged. > >> > > >> > So something should have changed somewhere within the newly installed > >> > Gromacs/VOTCA combination. > >> You can see what VOTCA has with: > >> csg_dump --top topol-rdf.tpr --cg dopc_cg-map11.xml > > > > > > OK, but its output does not contain any info about the exclusions. > > > >> > >> > >> Remember, if you are using a mapping file ("--cg" option) exclusions > >> from the topology file are ignored and the ones from the mapping file > >> are used instead. > > > > > > Alright, this must be the root of my problem! > > Apparently, in v1.3 csg_stat would take the exclusions from tpr, not xml, > > because I never provided any exclusion list in any xml file, but it did > work > > correctly! > This behavior hasn't changed for long time and was certainly the same > in v1.3 as well. > I am only reporting my experience: version 1.3 (installed on our cluster 8 Feb 2015) did not need any extra exclusion definitions (nrex) nor list and I had no problem obtaining correct distributions with topol-rdf.tpr only (the number provided by [moleculetype] is 10). But now, as of v1.41, the distributions correspond to exclusions starting with 1-4 intra-molecular pairs (the number provided by [moleculetype] is 3). Since I did not have any issue before, I did not care about generating exclusion lists in addition to tpr files. > When using a mapping file, exclusions get generated for each bond, > angle and dihedral interaction in the mapping file. > > Why are using a mapping file to begin with? > Good question! Well, the manual advises to use cg.inverse.map tag with a one-to-one mapping xml file for iterative methods even when optimising only non-bonded interactions (and I recall having problems without this tag!), so I have: <map> <DESC>Special mapping file for rdf calculations needed for bonded interactions</DESC> dopc_cg-map11.xml </map> and of course I do not do anything special about the csg_stat call in calc_rdf_generic.sh, so it does what it does. Andrey > Christoph > > > > I will look into this again. > > > > Andrey > > > >> > >> > >> Christoph > >> > > >> > Andrey > >> > > >> >> Christoph > >> >> > > >> >> > Andrey > >> >> > ----- > >> >> > $ grep rdf inverse.log > >> >> > ... > >> >> > cp_from_main_dir: 'grompp.mdp* topol.top* topol-rdf.tpr > table_??.xvg > >> >> > index.ndx* conf.gro*' > >> >> > cp_from_main_dir: 'grompp.mdp* topol.top* topol-rdf.tpr > table_??.xvg > >> >> > index.ndx* conf.gro*' > >> >> > cp_from_main_dir: 'grompp.mdp* topol.top* topol-rdf.tpr > table_??.xvg > >> >> > index.ndx* conf.gro*' > >> >> > cp_from_main_dir: 'grompp.mdp* topol.top* topol-rdf.tpr > table_??.xvg > >> >> > index.ndx* conf.gro*' > >> >> > cp_from_main_dir: 'grompp.mdp* topol.top* topol-rdf.tpr > table_??.xvg > >> >> > index.ndx* conf.gro*' > >> >> > ‘sim_0/topol-rdf.tpr’ -> ‘./topol-rdf.tpr’ > >> >> > Calculating rdfs with csg_stat using 4 tasks > >> >> > Calculating average rdfs and its errors for interaction TCH-TCH > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCH-TCO > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCO-TCO > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCH-PO > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCH-NH > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCO-PO > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction TCO-NH > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction PO-PO > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction PO-NH > >> >> > rdf calculation is already done > >> >> > Calculating average rdfs and its errors for interaction NH-NH > >> >> > > >> >> > > >> >> > > >> >> > On Sun, Jul 22, 2018 at 1:26 PM, Christoph Junghans > >> >> > <[email protected]> > >> >> > wrote: > >> >> >> > >> >> >> On Sat, Jul 21, 2018 at 9:01 PM, 'Andrey Brukhno' via votca > >> >> >> <[email protected]> wrote: > >> >> >> > Hello, > >> >> >> > > >> >> >> > I encountered a very strange (looks like legacy) issue with > >> >> >> > v1.4.1. > >> >> >> > In my earlier IBI iterations (with v1.3) I normally used special > >> >> >> > topol-rdf.tpr which was different from normal CG simulation > >> >> >> > topol.tpr > >> >> >> > by > >> >> >> > not > >> >> >> > including the intra-molecular pairs in the RDF stats (done by > >> >> >> > csg_stat). > >> >> >> > I > >> >> >> > wanted to make sure my previous results are reproduced by > v1.4.1, > >> >> >> > so > >> >> >> > copied > >> >> >> > all the initialisation files into a new directory and tried to > run > >> >> >> > once > >> >> >> > again. After the very first iteration I checked the obtained > RDFs > >> >> >> > and > >> >> >> > they > >> >> >> > were all wrong - corresponding to normal topol.tpr (including > >> >> >> > intra-pairs) > >> >> >> > and not topol-rdf.tpr (excluding intra-pairs)! > >> >> >> > > >> >> >> > I guess there might have been changes made to locations or names > >> >> >> > of > >> >> >> > tags > >> >> >> > in > >> >> >> > settings.xml between the VOTCA versions, but I could not > identify > >> >> >> > those > >> >> >> > changes by referring to the most recent manual (both v1.4.1 and > >> >> >> > development > >> >> >> > versions). > >> >> >> > > >> >> >> > Below is the relevant extract from my settings.xml file. > >> >> >> > > >> >> >> > Any clue, hint, advice are most welcome! > >> >> >> We haven't changes any tags in IBI since v1.4.1 and the special > >> >> >> topology for IBI can be set by cg.inverse.<sim_prog>.rdf.topol, > >> >> >> which > >> >> >> you did below. > >> >> >> You could try "grep csg_stat inverse.log" to figure out if the > right > >> >> >> topology got appended to the command line of csg_stat. > >> >> >> > >> >> >> Christoph > >> >> >> > > >> >> >> > Thanks in advance. > >> >> >> > Andrey > >> >> >> > > >> >> >> > ----------- > >> >> >> > <gromacs> > >> >> >> > <grompp> > >> >> >> > <bin>gmx grompp</bin> > >> >> >> > <!--opts></opts--> > >> >> >> > </grompp> > >> >> >> > <mdrun> > >> >> >> > <!--command>grmx-vtc-rerun.sh</command--> > >> >> >> > <!--command>mpirun -lsf mdrun_mpi</command--> > >> >> >> > <!--opts>-multidir sim_0 sim_1 sim_2 sim_3</opts--> > >> >> >> > <command>mpirun -n 4 mdrun_mpi</command> > >> >> >> > <opts>-multidir sim_0 sim_1 sim_2 sim_3 -tablep > table.xvg > >> >> >> > -tableb > >> >> >> > table_b?.xvg table_a?.xvg table_d?.xvg</opts> > >> >> >> > </mdrun> > >> >> >> > <!--mdrun><opts>-nt 2</opts></mdrun--> > >> >> >> > <!-- trash so many frames at the beginning --> > >> >> >> > <equi_time>0</equi_time> > >> >> >> > <!-- grid for table*.xvg !--> > >> >> >> > <table_bins>0.001</table_bins> > >> >> >> > <!-- cut the potential at this value (gromacs bug) --> > >> >> >> > <pot_max>10000</pot_max> > >> >> >> > <!-- extend the tables to this value --> > >> >> >> > <table_end>2.5</table_end> > >> >> >> > <initial_configuration>maindir</initial_configuration> > >> >> >> > <rdf> > >> >> >> > <block_length><DESC>Length of the block for the error > >> >> >> > analysis</DESC> > >> >> >> > 1000000 > >> >> >> > </block_length> > >> >> >> > <map><DESC>Special mapping file for rdf calculations > >> >> >> > needed > >> >> >> > for > >> >> >> > bonded interactions</DESC> > >> >> >> > dopc_cg-map11.xml > >> >> >> > </map> > >> >> >> > <topol><DESC> Gromacs binary topol (tpr) file to be used > >> >> >> > for > >> >> >> > csg_stat</DESC> > >> >> >> > topol-rdf.tpr > >> >> >> > </topol> > >> >> >> > <with_errors>yes<DESC>calculate error on the rdf: > >> >> >> > yes/no</DESC> > >> >> >> > </with_errors> > >> >> >> > </rdf> > >> >> >> > </gromacs> > >> >> >> > ----------- > >> >> >> > > >> >> >> > -- > >> >> >> > You received this message because you are subscribed to the > Google > >> >> >> > Groups > >> >> >> > "votca" group. > >> >> >> > To unsubscribe from this group and stop receiving emails from > it, > >> >> >> > send > >> >> >> > an > >> >> >> > email to [email protected]. > >> >> >> > To post to this group, send email to [email protected]. > >> >> >> > Visit this group at https://groups.google.com/group/votca. > >> >> >> > For more options, visit https://groups.google.com/d/optout. > >> >> >> > >> >> >> > >> >> >> > >> >> >> -- > >> >> >> Christoph Junghans > >> >> >> Web: http://www.compphys.de > >> >> >> > >> >> >> -- > >> >> >> You received this message because you are subscribed to a topic in > >> >> >> the > >> >> >> Google Groups "votca" group. > >> >> >> To unsubscribe from this topic, visit > >> >> >> https://groups.google.com/d/topic/votca/l42VuhKqQb4/unsubscribe. > >> >> >> To unsubscribe from this group and all its topics, send an email > to > >> >> >> [email protected]. > >> >> >> To post to this group, send email to [email protected]. > >> >> >> Visit this group at https://groups.google.com/group/votca. > >> >> >> For more options, visit https://groups.google.com/d/optout. > >> >> > > >> >> > > >> >> > -- > >> >> > You received this message because you are subscribed to the Google > >> >> > Groups > >> >> > "votca" group. > >> >> > To unsubscribe from this group and stop receiving emails from it, > >> >> > send > >> >> > an > >> >> > email to [email protected]. > >> >> > To post to this group, send email to [email protected]. > >> >> > Visit this group at https://groups.google.com/group/votca. > >> >> > For more options, visit https://groups.google.com/d/optout. > >> >> > >> >> > >> >> > >> >> -- > >> >> Christoph Junghans > >> >> Web: http://www.compphys.de > >> >> > >> >> -- > >> >> You received this message because you are subscribed to a topic in > the > >> >> Google Groups "votca" group. > >> >> To unsubscribe from this topic, visit > >> >> https://groups.google.com/d/topic/votca/l42VuhKqQb4/unsubscribe. > >> >> To unsubscribe from this group and all its topics, send an email to > >> >> [email protected]. > >> >> To post to this group, send email to [email protected]. > >> >> Visit this group at https://groups.google.com/group/votca. > >> >> For more options, visit https://groups.google.com/d/optout. > >> > > >> > > >> > -- > >> > You received this message because you are subscribed to the Google > >> > Groups > >> > "votca" group. > >> > To unsubscribe from this group and stop receiving emails from it, send > >> > an > >> > email to [email protected]. > >> > To post to this group, send email to [email protected]. > >> > Visit this group at https://groups.google.com/group/votca. > >> > For more options, visit https://groups.google.com/d/optout. > >> > >> > >> > >> -- > >> Christoph Junghans > >> Web: http://www.compphys.de > >> > >> -- > >> You received this message because you are subscribed to a topic in the > >> Google Groups "votca" group. > >> To unsubscribe from this topic, visit > >> https://groups.google.com/d/topic/votca/l42VuhKqQb4/unsubscribe. > >> To unsubscribe from this group and all its topics, send an email to > >> [email protected]. > >> To post to this group, send email to [email protected]. > >> Visit this group at https://groups.google.com/group/votca. > >> For more options, visit https://groups.google.com/d/optout. > > > > > > -- > > You received this message because you are subscribed to the Google Groups > > "votca" group. > > To unsubscribe from this group and stop receiving emails from it, send an > > email to [email protected]. > > To post to this group, send email to [email protected]. > > Visit this group at https://groups.google.com/group/votca. > > For more options, visit https://groups.google.com/d/optout. > > > > -- > Christoph Junghans > Web: http://www.compphys.de > > -- > You received this message because you are subscribed to a topic in the > Google Groups "votca" group. > To unsubscribe from this topic, visit https://groups.google.com/d/ > topic/votca/l42VuhKqQb4/unsubscribe. > To unsubscribe from this group and all its topics, send an email to > [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/votca. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "votca" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/votca. For more options, visit https://groups.google.com/d/optout.
