Dear Akram, just some remarks:
> My system is a molecule, so only the gamma point is used. For this, I used > Kgrid-Monkhorst-Pack block: > %block kgrid_Monkhorst_Pack > 1 0 0 0.5 > 0 1 0 0.5 > 0 0 1 0.5 > %endblock kgrid_Monkhorst_Pack strictly speaking this is NOT gamma point, but probably this won't make any difference for a molecule. > The system is not magnetic. Are you sure? How many electrons do you have? To me the behaviour you show below looks much like if your system attempts to occupy a narrow peak with a half-electron... > There is narrow peak in the density of states near the Fermi > energy (something like the one I attached) > I checked MeshCutoff Energy and increased it to 200 Ry but again diverge > SCF iteration! And when increase it more, program stop whit this error: > > InitMesh: Mesh cutoff (required, used) = 300.000 314.455 Ry This is not an error. Any anyway increasing MeshCutoff is not a sraight way to enforce the convergence. In my experience, you can almost always avoid the divergence by setting large enough (600 K ?) Electronic Temperature and setting small enough (0.0001 ? or less?) mixing parameter. But the convergence may not be fast. Moreover check if you don't have DM.NumberKick and corresponding mixing weight larger than it should be; it looks that something happens at some iterations, following a smooth development in between. Best regards Andrei Postnikov > > > > I send the iteration below: > > > > siesta: iscf Eharris(eV) E_KS(eV) FreeEng(eV) dDmax Ef(eV) > > siesta: 1 -3639.5345 -3608.1415 -3608.1415 1.9476 -4.4164 > > timer: Routine,Calls,Time,% = IterSCF 1 103.238 94.70 > > elaps: Routine,Calls,Wall,% = IterSCF 1 107.050 94.79 > > siesta: 2 -3890.1653 -3600.2437 -3600.4472 2.2513 0.3522 > > siesta: 3 -3695.5524 -3614.2091 -3614.2271 1.0335 -3.8205 > > siesta: 4 -3706.1134 -3613.8776 -3613.8776 0.9604 -3.3180 > > siesta: 5 -3746.9663 -3611.6947 -3611.6947 1.6417 -3.0051 > > siesta: 6 -3747.2714 -3611.6771 -3611.7568 1.6551 -3.0076 > > siesta: 7 -3747.5163 -3611.6605 -3611.7385 1.6654 -3.0099 > > siesta: 8 -3653.5162 -3620.5867 -3620.6633 0.8770 -4.2074 > > siesta: 9 -3644.7543 -3621.0979 -3621.1897 0.8945 -3.7655 > > siesta: 10 -3673.7327 -3619.8087 -3619.8753 0.8662 -3.8133 > > siesta: 11 -3672.9677 -3619.8175 -3619.8175 0.8632 -3.8537 > > siesta: 12 -3673.9518 -3619.8143 -3619.8143 0.8662 -3.8021 > > siesta: 13 -3667.9832 -3624.7232 -3624.7232 0.8548 -3.4625 > > siesta: 14 -3670.6356 -3623.4470 -3623.4470 0.8501 -3.4964 > > siesta: 15 -3671.9956 -3622.8412 -3622.8412 0.8477 -3.5056 > > siesta: 16 -3671.9725 -3625.6402 -3625.6402 0.8478 -3.0778 > > siesta: 17 -3664.6604 -3626.0750 -3626.0750 0.8880 -3.3645 > > siesta: 18 -3678.7139 -3625.7436 -3625.7436 0.8387 -2.6378 > > siesta: 19 -3681.1014 -3622.6742 -3622.6974 0.8428 -2.8061 > > siesta: 20 -3662.0336 -3635.4989 -3635.4990 0.7394 -2.6975 > > siesta: 21 -3655.5719 -3634.7390 -3634.7391 0.7538 -2.7755 > > siesta: 22 -3660.8795 -3635.8940 -3635.9095 0.7340 -2.6478 > > siesta: 23 -3659.8805 -3636.1779 -3636.1779 0.7347 -2.6106 > > siesta: 24 -3659.7537 -3634.8751 -3634.8753 0.7531 -2.7760 > > siesta: 25 -3651.2513 -3635.8139 -3635.8140 1.1318 -2.3884 > > siesta: 26 -3654.1426 -3634.3914 -3634.4502 0.7491 -2.8389 > > siesta: 27 -3653.2616 -3635.5981 -3635.6548 0.7568 -2.5686 > > siesta: 28 -3652.4034 -3637.1912 -3637.2555 1.3962 -2.0955 > > siesta: 29 -3655.8587 -3635.5721 -3635.5806 0.7348 -2.6678 > > siesta: 30 -3656.6067 -3635.4043 -3635.4200 0.7368 -2.6768 > > siesta: 31 -3647.5658 -3638.7697 -3638.7770 0.7151 -2.1191 > > siesta: 32 -3643.6205 -3637.8673 -3637.9482 0.6776 -2.4334 > > siesta: 33 -3641.3180 -3635.7890 -3635.8608 0.8582 -2.6286 > > siesta: 34 -3640.3337 -3636.1397 -3636.2475 1.0414 -2.5370 > > siesta: 35 -3640.6049 -3637.5830 -3637.6776 1.0250 -2.4267 > > siesta: 36 -3640.9347 -3638.2216 -3638.3261 0.8508 -2.4657 > > siesta: 37 -3641.5784 -3639.5665 -3639.6666 0.8512 -2.3566 > > siesta: 38 -3642.2852 -3639.0444 -3639.1329 0.5364 -2.6258 > > siesta: 39 -3641.6807 -3639.7086 -3639.7951 0.7085 -2.3911 > > siesta: 40 -3642.1041 -3639.5645 -3639.6490 0.5421 -2.5289 > > siesta: 41 -3641.9918 -3639.5672 -3639.6376 0.5420 -2.5027 > > siesta: 42 -3641.1518 -3638.8829 -3638.9536 1.4938 -2.3874 > > siesta: 43 -3641.8333 -3638.8825 -3638.8957 0.9692 -2.4500 > > siesta: 44 -3642.2133 -3638.1292 -3638.2189 0.5933 -2.5778 > > siesta: 45 -3642.4529 -3637.8025 -3637.8770 0.6024 -2.6787 > > siesta: 46 -3643.8639 -3639.3962 -3639.4653 0.5770 -2.7978 > > siesta: 47 -3643.1115 -3638.9763 -3639.0457 0.8764 -2.5604 > > siesta: 48 -3642.5625 -3639.0939 -3639.1872 0.5762 -2.5582 > > siesta: 49 -3642.1692 -3638.6502 -3638.7195 0.5724 -2.6051 > > siesta: 50 -3642.5710 -3638.5190 -3638.5894 0.5841 -2.6305 > > siesta: 51 -3642.0544 -3638.2051 -3638.2751 1.4403 -2.5436 > > siesta: 52 -3642.3849 -3638.9930 -3639.0739 1.5611 -2.4922 > > siesta: 53 -3642.8530 -3639.7468 -3639.7831 0.8088 -2.4522 > > siesta: 54 -3642.7870 -3639.6161 -3639.7144 0.5716 -2.4708 > > siesta: 55 -3642.7273 -3639.0012 -3639.0898 1.2220 -2.4938 > > siesta: 56 -3642.6019 -3638.9792 -3639.0503 1.1622 -2.4839 > > siesta: 57 -3642.7373 -3639.1627 -3639.2435 1.1319 -2.4836 > > siesta: 58 -3642.3415 -3638.7605 -3638.8418 0.9081 -2.5196 > > siesta: 59 -3642.3577 -3638.4266 -3638.5246 1.2098 -2.5247 > > siesta: 60 -3641.7507 -3637.9115 -3637.9956 1.1313 -2.5115 > > siesta: 61 -3642.5339 -3639.3943 -3639.4955 0.6408 -2.5266 > > .... and this has continuation. > > Regards: > > Akram > > --- On Mon, 5/3/10, [email protected] > <[email protected]> wrote: > > From: [email protected] <[email protected]> > Subject: Re: [SIESTA-L] About Convergence > To: [email protected] > Date: Monday, May 3, 2010, 11:53 AM > >> Hello, >> I have a problem about SCF iteration convergence. I did everything I >> guessed can help but it does not converge! I changed DM.MixingWeight >> (decreased it), DM.NumberPulay, DM.Tolerance even but nothing happened. >> I >> really do not know what should I do. >> Please someone help me. >> Regards; >> Akram >> > > Dear Akram, > it is difficult to help you without knowing the details. > > 1. What is the system you are calculating, bulk or molecule? > 2. Is the electronic structure reasonable? The bands due to > different atomic states situated where you expect them to be? > 3. Is the system magnetic, or should it be magnetic that you might forget? > 4. Do you have a narrow peak in the density of states near the Fermi > energy? Is the k-mesh sufficiently dense? > 5. As your calculation "does not converge", what is exactly going on? > Does your system swap between several competing states? Or, does it > go astray towards some unphysical results? > > Best regards > > Andrei Postnikov > > > > > >
