Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Hi I am doing calculations for Mn using lda+U.I GOT the same problem.but when I do with lda then there was no problem. I could not understand. Plz help me.if anybody have mn.psf which are compatible with lda+u.plz send me. Thanks. On 21 Jan 2015 21:22, "Seyed Mohammad Tabatabaei" wrote: > Hi, > > I had a similar problem and settled it by changing the PAO.SplitNorm to > the minimum split norm parameter and it worked. However, I would like to > know about the physical meaning much more. > > Bests regards, > > On Wed, Jan 21, 2015 at 5:32 PM, gagandeep Kaur > wrote: > >> Thnx a ton fr d reply..i tried reading it..simply shud i change >> PAO.SplitNorm= 0.016412 >> >> >> Sent from Samsung Mobile >> >> >> Original message >> From: Nick Papior Andersen >> Date:21/01/2015 18:29 (GMT+05:30) >> To: siesta-l@uam.es >> Subject: Re: [SIESTA-L] please help See manual for new split options >> ERROR STOP from Node: 0 >> >> >> >> 2015-01-21 12:48 GMT+00:00 Gagandeep kaur : >> >>> Dear SIESTA users >>> i am doing calculations on clusters adsorbed on 32 atom graphene >>> sheet..fdf file is attached herewith..i encounter with following >>> problem..kindly send me the solution.. >>> >>> Gagandeep Kaur >>> Research Scholar >>> Panjab University >>> Chandigarh >>> India >>> >>> >>> >>> We have tried to run command "mpirun -np 4 siesta < C32T.fdf" >>> with 1, 2 and 4 number of processes. >>> >>>With each run the below mentioned error came: >>> >>> ## >>> >>> WARNING: Minimum split_norm parameter: 0.16412. Will not be able to >>> generate orbital with split_norm = 0.15000 >>> >>> See manual for new split options >>> ERROR STOP from Node:0 >>> >> This error message is quite specific, the split norm parameter is set to >> 0.15, however, siesta calculates that the minimum required split norm >> parameter needs to be (at least) 0.16412. >> Hence you should search in the manual on how to change the split norm >> parameter. >> The line: >> "See manual for new split options" >> is pretty exact in this case :) >> >> Enjoy the read. >> >> >> >> -- >> Kind regards Nick >> > >
Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Hi, I had a similar problem and settled it by changing the PAO.SplitNorm to the minimum split norm parameter and it worked. However, I would like to know about the physical meaning much more. Bests regards, On Wed, Jan 21, 2015 at 5:32 PM, gagandeep Kaur wrote: > Thnx a ton fr d reply..i tried reading it..simply shud i change > PAO.SplitNorm= 0.016412 > > > Sent from Samsung Mobile > > > Original message > From: Nick Papior Andersen > Date:21/01/2015 18:29 (GMT+05:30) > To: siesta-l@uam.es > Subject: Re: [SIESTA-L] please help See manual for new split options ERROR > STOP from Node: 0 > > > > 2015-01-21 12:48 GMT+00:00 Gagandeep kaur : > >> Dear SIESTA users >> i am doing calculations on clusters adsorbed on 32 atom graphene >> sheet..fdf file is attached herewith..i encounter with following >> problem..kindly send me the solution.. >> >> Gagandeep Kaur >> Research Scholar >> Panjab University >> Chandigarh >> India >> >> >> >> We have tried to run command "mpirun -np 4 siesta < C32T.fdf" >> with 1, 2 and 4 number of processes. >> >>With each run the below mentioned error came: >> >> ## >> >> WARNING: Minimum split_norm parameter: 0.16412. Will not be able to >> generate orbital with split_norm = 0.15000 >> >> See manual for new split options >> ERROR STOP from Node:0 >> > This error message is quite specific, the split norm parameter is set to > 0.15, however, siesta calculates that the minimum required split norm > parameter needs to be (at least) 0.16412. > Hence you should search in the manual on how to change the split norm > parameter. > The line: > "See manual for new split options" > is pretty exact in this case :) > > Enjoy the read. > > > > -- > Kind regards Nick >
Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Thnx a ton fr d reply..i tried reading it..simply shud i change PAO.SplitNorm= 0.016412 Sent from Samsung Mobile Original message From: Nick Papior Andersen Date:21/01/2015 18:29 (GMT+05:30) To: siesta-l@uam.es Subject: Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0 2015-01-21 12:48 GMT+00:00 Gagandeep kaur : Dear SIESTA users i am doing calculations on clusters adsorbed on 32 atom graphene sheet..fdf file is attached herewith..i encounter with following problem..kindly send me the solution.. Gagandeep Kaur Research Scholar Panjab University Chandigarh India We have tried to run command "mpirun -np 4 siesta < C32T.fdf" with 1, 2 and 4 number of processes. With each run the below mentioned error came: ## WARNING: Minimum split_norm parameter: 0.16412. Will not be able to generate orbital with split_norm = 0.15000 See manual for new split options ERROR STOP from Node:0 This error message is quite specific, the split norm parameter is set to 0.15, however, siesta calculates that the minimum required split norm parameter needs to be (at least) 0.16412. Hence you should search in the manual on how to change the split norm parameter. The line: "See manual for new split options" is pretty exact in this case :) Enjoy the read. -- Kind regards Nick
Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Thnx fr d reply bt i hv nt added energy shift in fdf attached Sent from Samsung Mobile Original message From: Walter Huang Date:21/01/2015 18:31 (GMT+05:30) To: siesta-l@uam.es Cc: siesta-l Subject: Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0 Hi, A very simple way to avoid this is to decrease the value of energy_shift. Best regards, Walter On Jan 21, 2015, at 08:50 PM, Gagandeep kaur wrote: Dear SIESTA users i am doing calculations on clusters adsorbed on 32 atom graphene sheet..fdf file is attached herewith..i encounter with following problem..kindly send me the solution.. Gagandeep Kaur Research Scholar Panjab University Chandigarh India We have tried to run command "mpirun -np 4 siesta < C32T.fdf" with 1, 2 and 4 number of processes. With each run the below mentioned error came: ## WARNING: Minimum split_norm parameter: 0.16412. Will not be able to generate orbital with split_norm = 0.15000 See manual for new split options ERROR STOP from Node: 0
Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
2015-01-21 12:48 GMT+00:00 Gagandeep kaur : > Dear SIESTA users > i am doing calculations on clusters adsorbed on 32 atom graphene > sheet..fdf file is attached herewith..i encounter with following > problem..kindly send me the solution.. > > Gagandeep Kaur > Research Scholar > Panjab University > Chandigarh > India > > > > We have tried to run command "mpirun -np 4 siesta < C32T.fdf" > with 1, 2 and 4 number of processes. > >With each run the below mentioned error came: > > ## > > WARNING: Minimum split_norm parameter: 0.16412. Will not be able to > generate orbital with split_norm = 0.15000 > > See manual for new split options > ERROR STOP from Node:0 > This error message is quite specific, the split norm parameter is set to 0.15, however, siesta calculates that the minimum required split norm parameter needs to be (at least) 0.16412. Hence you should search in the manual on how to change the split norm parameter. The line: "See manual for new split options" is pretty exact in this case :) Enjoy the read. -- Kind regards Nick
Re: [SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Hi,A very simple way to avoid this is to decrease the value of energy_shift.Best regards,WalterOn Jan 21, 2015, at 08:50 PM, Gagandeep kaur wrote:Dear SIESTA usersi am doing calculations on clusters adsorbed on 32 atom graphenesheet..fdf file is attached herewith..i encounter with followingproblem..kindly send me the solution..Gagandeep KaurResearch ScholarPanjab UniversityChandigarhIndia We have tried to run command "mpirun -np 4 siesta < C32T.fdf"with 1, 2 and 4 number of processes. With each run the below mentioned error came:##WARNING: Minimum split_norm parameter: 0.16412. Will not be able togenerate orbital with split_norm = 0.15000See manual for new split optionsERROR STOP from Node: 0 C32T.fdf Description: Binary data
[SIESTA-L] please help See manual for new split options ERROR STOP from Node: 0
Dear SIESTA users i am doing calculations on clusters adsorbed on 32 atom graphene sheet..fdf file is attached herewith..i encounter with following problem..kindly send me the solution.. Gagandeep Kaur Research Scholar Panjab University Chandigarh India We have tried to run command "mpirun -np 4 siesta < C32T.fdf" with 1, 2 and 4 number of processes. With each run the below mentioned error came: ## WARNING: Minimum split_norm parameter: 0.16412. Will not be able to generate orbital with split_norm = 0.15000 See manual for new split options ERROR STOP from Node:0 C32T.fdf Description: Binary data
[SIESTA-L] Please help me in geometry optimization
Dear Siesta Users, I want to do a geometry optimization. I preprepared the input file for running siesta. However something is confused me and I will appreciate anyone who can help me to understand these points: 1- I choose CG type of run and I select number of CG steps=100, How it is possible to determine that it is enough for such an optimization? 2- Is Siesta use energy convergence for its calculation? How can enter this in the input file? 3- If I want to study the interaction between two species, is the primary position important for siesta? Best regards, E. Shakerzadeh
Re: [SIESTA-L] Please help
Hi, THank you very much. I spent 2 days already on this simulation. I will be more careful next time to write the correct number of atoms. REgards, Varrun Ashok On 18 April 2011 16:03, Natalia Martsinovich wrote: > Hello Varrun, > > Your input and output files seem OK, but the optimisation is not yet > finished: you can check that the forces are decreasing (but are still > large), and energies (E_KS(eV)) are converging, but are still far from > being converged. 26 optimisation steps is clearly not enough yet. > > Few more points: the xyz file has 171 atoms, but NumberOfAtoms is 168, > so your calculation runs with only one Ti atom, not 4. > Also, I advise to keep the cell volume fixed, to save computational > time and not allow the cell to become too small: > MD.VariableCell F > > Best regards, > > Natalia > > On 4/16/11, Varrun Ashok wrote: > > Hi all, > > I am very new to siesta and DFT calculations. I need your support and > help. > > I am running a fdf file comprising a system consisting of a triglyceride > > molecule and a 4 atom Ti tetrahedral cluster near a double of the > molecule. > > The aim is to check how strongly the cluster interacts with double bond. > The > > system is not converging. The Max atomic forces values are erratic. I > have > > attached a copy of the fdf file, the .xyz file and the .out file and also > a > > 3D image of the molecule+Ti system. Kindly provide recommendations. > > Regards, > > Varrun Ashok > > >
Re: [SIESTA-L] Please help
Hello Varrun, Your input and output files seem OK, but the optimisation is not yet finished: you can check that the forces are decreasing (but are still large), and energies (E_KS(eV)) are converging, but are still far from being converged. 26 optimisation steps is clearly not enough yet. Few more points: the xyz file has 171 atoms, but NumberOfAtoms is 168, so your calculation runs with only one Ti atom, not 4. Also, I advise to keep the cell volume fixed, to save computational time and not allow the cell to become too small: MD.VariableCell F Best regards, Natalia On 4/16/11, Varrun Ashok wrote: > Hi all, > I am very new to siesta and DFT calculations. I need your support and help. > I am running a fdf file comprising a system consisting of a triglyceride > molecule and a 4 atom Ti tetrahedral cluster near a double of the molecule. > The aim is to check how strongly the cluster interacts with double bond. The > system is not converging. The Max atomic forces values are erratic. I have > attached a copy of the fdf file, the .xyz file and the .out file and also a > 3D image of the molecule+Ti system. Kindly provide recommendations. > Regards, > Varrun Ashok >
Re: [SIESTA-L] Please, help with LDOS 2D plot (arbitrary plane)
Ruslan, Most probably by using XCrysDen you will be able to plot what you want. I know that in XCrysDen you can, for example, define a plane by choosing three atoms and it will plot th charge density (thus the LDOS as well), but I'm not sure of how to do it exactly. Also, check Andrei Postnikov's utilities, they will most likely contain something that will do what you want :) Marcos On Thu, Jul 8, 2010 at 3:19 PM, Руслан Жачук wrote: > Dear Siesta users, > > > I need to plot in 2D LDOS of my system integrated between two energy levels > in some specific plane which is not generally parallel or perpendicular to > any of x,y,z axis. > The plstm utility is good, but to use it the plane of plot must be > perpendicular to z-axis. > > > The denchar utility works with arbitrary planes, but for charge density and > wavefunctions only, not LDOS. > Do you know the way how LDOS can be plotted in 2D, using arbitrary plane? > Can denchar be adapted so that it can plot LDOS (RHO and LDOS files have the > same format)? > > > Of course there is always a way to rotate coordinate system so that the > plane of interest will be perpendicular to z-axis, but this solution seems > to be cumbersome... > > > Regards > Ruslan Zhachuk >
Re: Posible SPAM: Re: [SIESTA-L] Please, help with LDOS 2D plot (arbitrary plane)
I need to visualize 2D plot of LDOS, and also save it, because I need to compare isolines of similar systems. Not sure it is possible in XCrysden. Now I think Util/Contour in Siesta package would be the right choice. I will try this. Ruslan 2010/7/9 > > Ruslan, > > RHO and LDOS files have exactly the same format > > but as far as I understand (not sure 100%) Denchar does not read them to > > plot the charge density, it uses the wavefunctions directly. > > But no, Ruslan - > denchar is perfectly suited to plot either wavefunctions or charge density, > there is switch in it about which one you want to plot. > And it perfectly allows to make 2-dim plots. > However, I think it adds file extensions by default, > so there might be no way to plot LDOS, > unless you rename your LDOS file into RHO for a couple of minutes. > > > Try /Util/Contrib/APostnikov changing the direction of the unit cell > > vectors and plotting with xcrysden > > this should work, but isolines options in XCrysden are not so flexible > as e.g. in gnuplot. > > Good luck > > Andrei Postnikov > > > or > > Util/Contour (not sure it does what you need but looks like that was the > > point of the program) > > > > or > > /home/voznyy/SIESTA/SRC/siesta-3.0-rc1/Util/Grid/grid2cube > > and plot the 3D cube file with > > http://www.geocities.jp/kmo_mma/crystal/en/vesta.html > > it can then cut through any plane > > > > Alexander. > > > >
Posible SPAM: Re: [SIESTA-L] Please, help with LDOS 2D plot (arbitrary plane)
> Ruslan, > RHO and LDOS files have exactly the same format > but as far as I understand (not sure 100%) Denchar does not read them to > plot the charge density, it uses the wavefunctions directly. But no, Ruslan - denchar is perfectly suited to plot either wavefunctions or charge density, there is switch in it about which one you want to plot. And it perfectly allows to make 2-dim plots. However, I think it adds file extensions by default, so there might be no way to plot LDOS, unless you rename your LDOS file into RHO for a couple of minutes. > Try /Util/Contrib/APostnikov changing the direction of the unit cell > vectors and plotting with xcrysden this should work, but isolines options in XCrysden are not so flexible as e.g. in gnuplot. Good luck Andrei Postnikov > or > Util/Contour (not sure it does what you need but looks like that was the > point of the program) > > or > /home/voznyy/SIESTA/SRC/siesta-3.0-rc1/Util/Grid/grid2cube > and plot the 3D cube file with > http://www.geocities.jp/kmo_mma/crystal/en/vesta.html > it can then cut through any plane > > Alexander.
Posible SPAM: Re: [SIESTA-L] Please, help with LDOS 2D plot (arbitrary plane)
Ruslan, RHO and LDOS files have exactly the same format but as far as I understand (not sure 100%) Denchar does not read them to plot the charge density, it uses the wavefunctions directly. Try /Util/Contrib/APostnikov changing the direction of the unit cell vectors and plotting with xcrysden or Util/Contour (not sure it does what you need but looks like that was the point of the program) or /home/voznyy/SIESTA/SRC/siesta-3.0-rc1/Util/Grid/grid2cube and plot the 3D cube file with http://www.geocities.jp/kmo_mma/crystal/en/vesta.html it can then cut through any plane Alexander.
[SIESTA-L] Please, help with LDOS 2D plot (arbitrary plane)
Dear Siesta users, I need to plot in 2D LDOS of my system integrated between two energy levels in some specific plane which is not generally parallel or perpendicular to any of x,y,z axis. The plstm utility is good, but to use it the plane of plot must be perpendicular to z-axis. The denchar utility works with arbitrary planes, but for charge density and wavefunctions only, not LDOS. Do you know the way how LDOS can be plotted in 2D, using arbitrary plane? Can denchar be adapted so that it can plot LDOS (RHO and LDOS files have the same format)? Of course there is always a way to rotate coordinate system so that the plane of interest will be perpendicular to z-axis, but this solution seems to be cumbersome... Regards Ruslan Zhachuk
Re: [SIESTA-L] Please help
Dear Akram, < I expected the band gap in DOS decreased strongly, but nothing happened overall! Why you have such expectation? have you any experimental data which support your expectation? if yes have you considerd the correct substitution position for your impurities? Regards On Fri, May 7, 2010 at 12:08 PM, Akram Keyhannezhad < akram.keyhannez...@yahoo.com> wrote: > Dear siesta users; > I calculated polyacetylene chain DOS in siesta. Then I constructed a .fdf > file for polyacetylene MOLECULE and got nearly reasonable results. Now, > I added an impurity (like Iodine, Li, N, Na). I expected the band gap in DOS > decreased strongly, but nothing happened overall! > Where is the problem in your idea? Can it because of DOS block or maybe > another blocks in .fdf? I mean can I improve result by making some changes > in fdf or I must check again system definition? > Thanks alot for your help before. > Best regards; > Akram > > -- Professor M. R. Benam, PhD Head of Physics Department , Payame-Noor University, Mashhad, Iran.
[SIESTA-L] Please help
Dear siesta users; I calculated polyacetylene chain DOS in siesta. Then I constructed a .fdf file for polyacetylene MOLECULE and got nearly reasonable results. Now, I added an impurity (like Iodine, Li, N, Na). I expected the band gap in DOS decreased strongly, but nothing happened overall! Where is the problem in your idea? Can it because of DOS block or maybe another blocks in .fdf? I mean can I improve result by making some changes in fdf or I must check again system definition? Thanks alot for your help before. Best regards; Akram
Re: [SIESTA-L] Please, help with charge density
Dear Andrei and Ruslan Just a quick note on your discussion: apost...@uni-osnabrueck.de wrote: Dear Ruslan: No it is not in the manual; however in the file m_iorho.F, subroutine write_rho, you could theoretically search for the write(iu) statements and replace them by formatted ones, or provide the switch. Unfortunately, I am not sure that this is allowed by the Siesta licence. Yes, it is. You can modify siesta in whatever way you want if you have a license and it is for your own consumption. The license limits the redistribution of the code, modified or not. If you want to export your modification(s) to others, that is another matter, although there are ways contemplated in the license. But feel free to change things for your own (that is one of the benefits of having the source). Emilio -- Emilio Artacho Department of Earth Sciences, University of Cambridge Downing Street, Cambridge CB2 3EQ, UK Tel. (+44/0) 1223 333480, Fax (+44/0) 1223 333450 emi...@esc.cam.ac.uk, http://www.esc.cam.ac.uk/~emilio
Re: [SIESTA-L] Please, help with charge density
>> > Finally I will need charge density integrated along Z direction, i.e. >> > n(x,y). >> >> What useful meaning would it have, integrated along Z??? - just >> wondering... >> > > I want to plot in 2D redistribution of charges when an atom is placed on > the > surface. > So I will calculate charge density of (surface + atom) integrated in Z: > n(x,y) > Then calculate it just for (surface) with atoms frozen in same positions: > m(x,y) > Then calculate it just for (atom): p(x,y) > And finally I will need n(x,y)-m(x,y)-p(x,y) > > > >> > Is it possible to do this by using rho2xsf tool and specifing Z >> division >> > =1??? >> > I suppose the charge density will be integrated along Z??? >> >> No it won't integrate, it will interpolate the value in the plane. >> > > Hmmm. So I will need to make *.XSF with 216x216x216 values and integrate > them manually? Dear Ruslan, there are other ways to integrate than manually. But if you don't want to touch the code (any code), make 216 (x,y) scans, merge them in a single .XSF file and sum up within the XCrySDen. In fact for your purpose - I think you are interested in a rough estimate - you do not need to make exactly 216 scans, a dozen will probably suffice... Best regards Andrei Postnikov
Re: [SIESTA-L] Please, help with charge density
> > Finally I will need charge density integrated along Z direction, i.e. > > n(x,y). > > What useful meaning would it have, integrated along Z??? - just > wondering... > I want to plot in 2D redistribution of charges when an atom is placed on the surface. So I will calculate charge density of (surface + atom) integrated in Z: n(x,y) Then calculate it just for (surface) with atoms frozen in same positions: m(x,y) Then calculate it just for (atom): p(x,y) And finally I will need n(x,y)-m(x,y)-p(x,y) > > Is it possible to do this by using rho2xsf tool and specifing Z division > > =1??? > > I suppose the charge density will be integrated along Z??? > > No it won't integrate, it will interpolate the value in the plane. > Hmmm. So I will need to make *.XSF with 216x216x216 values and integrate them manually? Kind regards Ruslan
Re: [SIESTA-L] Please, help with charge density
> Not trivial at all and not described in the manual :( This is for users who know what they are doing :) > But ok, if formatted *.RHO files are so huge they are not useful. What do you mean? I thought, if you really need this information you won't care..? After all, the hugeness of this file does also depend in its format. > Finally I will need charge density integrated along Z direction, i.e. > n(x,y). What useful meaning would it have, integrated along Z??? - just wondering... > Is it possible to do this by using rho2xsf tool and specifing Z division > =1??? > I suppose the charge density will be integrated along Z??? No it won't integrate, it will interpolate the value in the plane. > And I would prefer to keep the same grid in X, Y which was used internally > in Siesta. This is your good right > Is it correct from below that each spanning vector of the unit cell is > devided in 216 parts to make the grid??? > outcell: Cell vector modules (Ang) : 27.174106 27.174107 25.00 > outcell: Cell angles (23,13,12) (deg): 90. 90. 60. > outcell: Cell volume (Ang**3): 15987.5228 > > InitMesh: MESH = 216 x 216 x 216 =10077696 > InitMesh: Mesh cutoff (required, used) = 200.000 206.314 Ry Seems so. It is funny that such different cell vectors as 27.14 and 25.00 are divided with exactly equal number of mesh points, but so it might be, due to constraints of the FFT mesh. Best regards Andrei Postnikov
Re: [SIESTA-L] Please, help with charge density
Ok, I understand. But now it seems to be simpler to do with rho2xsf. Finally I will need charge density integrated along Z direction, i.e. n(x,y). Is it possible to do this by using rho2xsf tool and specifing Z division =1??? I suppose the charge density will be integrated along Z??? And I would prefer to keep the same grid in X, Y which was used internally in Siesta. Is it correct from below that each spanning vector of the unit cell is devided in 216 parts to make the grid??? outcell: Cell vector modules (Ang) : 27.174106 27.174107 25.00 outcell: Cell angles (23,13,12) (deg): 90. 90. 60. outcell: Cell volume (Ang**3): 15987.5228 InitMesh: MESH = 216 x 216 x 216 =10077696 InitMesh: Mesh cutoff (required, used) = 200.000 206.314 Ry Kind regards Ruslan Zhachuk Dear Ruslan: > No it is not in the manual; however in the file m_iorho.F, > subroutine write_rho, > you could theoretically search for the write(iu) statements > and replace them by formatted ones, or provide the switch. > Unfortunately, I am not sure that this is allowed by the Siesta licence. > > Best regards > > Andrei Postnikov > > >
Re: [SIESTA-L] Please, help with charge density
Not trivial at all and not described in the manual :( But ok, if formatted *.RHO files are so huge they are not useful. Finally I will need charge density integrated along Z direction, i.e. n(x,y). Is it possible to do this by using rho2xsf tool and specifing Z division =1??? I suppose the charge density will be integrated along Z??? And I would prefer to keep the same grid in X, Y which was used internally in Siesta. Is it correct from below that each spanning vector of the unit cell is devided in 216 parts to make the grid??? outcell: Cell vector modules (Ang) : 27.174106 27.174107 25.00 outcell: Cell angles (23,13,12) (deg): 90. 90. 60. outcell: Cell volume (Ang**3): 15987.5228 InitMesh: MESH = 216 x 216 x 216 =10077696 InitMesh: Mesh cutoff (required, used) = 200.000 206.314 Ry Kind regards Ruslan Zhachuk 2009/12/11 Marcos Veríssimo Alves > In the 3.0-b version, I'd guess you'd go to the file m_iorho.F, change line > 37 from > > character(len=*), parameter :: fform = "unformatted" > > to > > character(len=*), parameter :: fform = "formatted" > > > and re-build siesta. Bear in mind that formatted files are much larger than > unformatted ones (meaning huge). > > Marcos > > On Fri, Dec 11, 2009 at 12:22 PM, Руслан Жачук wrote: > >> >> Dear Andrei, >> >> thank you for your reply. With rho2xsf I agree, it works. >> But how to switch Siesta to formatted output (formatted RHO)?? >> I don't see such statements in Siesta-2.0.2 manual. >> >> Regards >> Ruslan Zhachuk >> >> >>> > >>> > With 3) it is in principle possible to make *.xsf file and then extract >>> > values. The problem is that rho2xsf utility is interactive and hence is >>> > not appropriate for work with big files (446 atoms) on cluster >>> remotely. >>> >>> Dear Ruslan: >>> Why? You know the questions the rho2xsf will ask you. >>> They are not about every one of your 446 atoms, but about the box origin, >>> size, and number of divisions. You can prepare a (short) file >>> with answers and feed in in to rho2xsf on your precious cluster, >>> rho2xsf < INP > OUT >>> Or, you hack the source and change it to your needs. >>> Especially if you eventually need the result in a different format. >>> (But please give it a different name :-) >>> Or - if you need the RHO data exactly on the Siesta mesh over the cell, >>> it might be easier to change the write statement in Siesta >>> to a formatted output >>> >>> Best regards >>> >>> Andrei Postnikov >>> >>> >>> >> >
Re: [SIESTA-L] Please, help with charge density
> Dear Andrei, > > thank you for your reply. With rho2xsf I agree, it works. > But how to switch Siesta to formatted output (formatted RHO)?? > I don't see such statements in Siesta-2.0.2 manual. Dear Ruslan: No it is not in the manual; however in the file m_iorho.F, subroutine write_rho, you could theoretically search for the write(iu) statements and replace them by formatted ones, or provide the switch. Unfortunately, I am not sure that this is allowed by the Siesta licence. Best regards Andrei Postnikov
Re: [SIESTA-L] Please, help with charge density
In the 3.0-b version, I'd guess you'd go to the file m_iorho.F, change line 37 from character(len=*), parameter :: fform = "unformatted" to character(len=*), parameter :: fform = "formatted" and re-build siesta. Bear in mind that formatted files are much larger than unformatted ones (meaning huge). Marcos On Fri, Dec 11, 2009 at 12:22 PM, Руслан Жачук wrote: > > Dear Andrei, > > thank you for your reply. With rho2xsf I agree, it works. > But how to switch Siesta to formatted output (formatted RHO)?? > I don't see such statements in Siesta-2.0.2 manual. > > Regards > Ruslan Zhachuk > > >> > >> > With 3) it is in principle possible to make *.xsf file and then extract >> > values. The problem is that rho2xsf utility is interactive and hence is >> > not appropriate for work with big files (446 atoms) on cluster remotely. >> >> Dear Ruslan: >> Why? You know the questions the rho2xsf will ask you. >> They are not about every one of your 446 atoms, but about the box origin, >> size, and number of divisions. You can prepare a (short) file >> with answers and feed in in to rho2xsf on your precious cluster, >> rho2xsf < INP > OUT >> Or, you hack the source and change it to your needs. >> Especially if you eventually need the result in a different format. >> (But please give it a different name :-) >> Or - if you need the RHO data exactly on the Siesta mesh over the cell, >> it might be easier to change the write statement in Siesta >> to a formatted output >> >> Best regards >> >> Andrei Postnikov >> >> >> >
Re: [SIESTA-L] Please, help with charge density
Dear Andrei, thank you for your reply. With rho2xsf I agree, it works. But how to switch Siesta to formatted output (formatted RHO)?? I don't see such statements in Siesta-2.0.2 manual. Regards Ruslan Zhachuk > > > > With 3) it is in principle possible to make *.xsf file and then extract > > values. The problem is that rho2xsf utility is interactive and hence is > > not appropriate for work with big files (446 atoms) on cluster remotely. > > Dear Ruslan: > Why? You know the questions the rho2xsf will ask you. > They are not about every one of your 446 atoms, but about the box origin, > size, and number of divisions. You can prepare a (short) file > with answers and feed in in to rho2xsf on your precious cluster, > rho2xsf < INP > OUT > Or, you hack the source and change it to your needs. > Especially if you eventually need the result in a different format. > (But please give it a different name :-) > Or - if you need the RHO data exactly on the Siesta mesh over the cell, > it might be easier to change the write statement in Siesta > to a formatted output > > Best regards > > Andrei Postnikov > > >
Re: [SIESTA-L] Please, help with charge density
> Dear siesta users, > > I need charge density in format "x y z (density_value)" on the grid used > by > Siesta (which is set by MeshCutoff parameter). > From what I learnt the charge density can be processed by 4 utilities: > > 1) plrho (isosurfaces) > 2) WSxM (isosurfaces) > 3) rho2xsf > 4) denchar > > 1) and 2) actually don't give access to grid values. > > With 3) it is in principle possible to make *.xsf file and then extract > values. The problem is that rho2xsf utility is interactive and hence is > not appropriate for work with big files (446 atoms) on cluster remotely. Dear Ruslan: Why? You know the questions the rho2xsf will ask you. They are not about every one of your 446 atoms, but about the box origin, size, and number of divisions. You can prepare a (short) file with answers and feed in in to rho2xsf on your precious cluster, rho2xsf < INP > OUT Or, you hack the source and change it to your needs. Especially if you eventually need the result in a different format. (But please give it a different name :-) Or - if you need the RHO data exactly on the Siesta mesh over the cell, it might be easier to change the write statement in Siesta to a formatted output Best regards Andrei Postnikov
[SIESTA-L] Please, help with charge density
Dear siesta users, I need charge density in format "x y z (density_value)" on the grid used by Siesta (which is set by MeshCutoff parameter). >From what I learnt the charge density can be processed by 4 utilities: 1) plrho (isosurfaces) 2) WSxM (isosurfaces) 3) rho2xsf 4) denchar 1) and 2) actually don't give access to grid values. With 3) it is in principle possible to make *.xsf file and then extract values. The problem is that rho2xsf utility is interactive and hence is not appropriate for work with big files (446 atoms) on cluster remotely. With 4) it is possible to make *.cube file and then extract values. But denchar uses only rectangular cell (which is not the case for me) and also grid points is different from what Siesta used internally. Please, help Ruslan Zhachuk