Manual



Installation guide

Binary of KPACK is distributed in a tar.gz file, which can be extracted by,

$ tar xvfz kpack-2014.7-bin.tar.gz


To run KPACK

The output is printed to the screen, and therefore, a normal execution is performed as below. 

$ kpack input_fileoutput_file &

The progress can be checked by, 

$ tail –f output_file


Input structure

Fortran NAMELIST style is used to construct the input file. There are 7 NAMELIST groups that must be present in the input even though some of these groups are not used in the selected execution (so leave them blank), and the groups are as follows.

kpack : decides the method of calculation
ao : specify geometry, basis set
geom       : geometry scan
scf : SCF variables
trans        : post-HF molecular spinor transformation 
ci : configuration interaction variables
perm : permute the SCF spinors before CASSCF
casscf : CASSCF module

A NAMELIST group is presented in the input as,

&group
 var_int=3 !integer
 var_real=1.0 !real (also 1.0d0 is possible)
 var_log=.true. !logical
 var_str=”ECPCE92SO” !character string
&end


Input group: kpack

 Data typeDescriptionDefault 
method character(len=4) Specifies SCF procedure
 "krhf" or "kuhf"
 "krhf" 
 mp2 logical KR/UMP2 calculation
 trans group must be defined
 .false. 
 ci logical KR/UCI calculation
 trans and ci groups must be defined
 .false.
 casscf logical KRCASSCF calcualtion
cas group must be defined
 .false.

Sample1: KUCI calculation
&kpack
 method=“kuhf”
 ci=.true.
&end

Sample2: KRCASSCF calculation
&kpack
 method=”krhf”
 casscf=.true.
&end


Input group: ao

memgb real
                maximum memory that can be allocated during AO integration
                Choice: 0.0-Maximum physical memory
Default: 1.0d0 (i.e. 1GB)

coord derived data type [character(len=4), 3*real]
nuclear Cartesian coordinate of molecule
Choice: atomic symbol, x, y, z

basis derived data type 
[character(len=4), 2*character(len=50)]
basis set and SOREP information
Choice: atomic symbol, basis set, SOREP
SOREP part should be set to “none” if SOREP is not available
                The basis sets and SOREPs should be listed at the end of input file.
                Note that there should be no identical strings.
                In other words, the same name of basis or SOREP cannot be used
                for more than one element.
                e.g., say cc-pVDZ is used for C and H, then use "cc-pVDZ_H" for hydrogen
                and "cc-pVDZ_C" for carbon.
                See Sample1 for the basis and SOREP formats used in KPACK.
                Nonrelativistic basis sets may be obtained from EMSL Basis Set Exchange,
                and SOREPs and accompanying basis sets from the websites given in here.

gaussian character(len=1)
basis set type: cartessian or spherical
Choice: “C” or “S”

Sample1: TlH (r=1.8 Å)
&ao
 coord =
  “Tl” 0.0 0.0 0.0
  “H”  0.0 0.0 1.8
 basis =
  “Tl” “ECPCE68_DZ” “ECPCE68”
  “H”  “cc-pVDZ” “none”
 gaussian = “S”
&end
...
ECPCE68                                        Name of SOREP
4 4                                                   Highest angular momentum (G in this case) for AREP and SO potentials, respectively
6                                                      block (AREP)
2       1.188500       -1.305280
2       2.826400      -11.615490
2       8.247200      -44.717980
2      20.529000     -130.115524
2      70.763200     -340.590473
1     223.323400      -53.444254
7                                                       S-G block (AREP)
2        .985000       44.021089
2       1.135800     -142.719578
2       1.489600      282.763667
2       2.107900     -306.241150
2       2.925500      186.004451
1      10.644300       55.996380
0      34.876400        6.371788
8                                                        P-G block (AREP)
2        .290400         .115591
2        .920100      -54.439704
2       1.084500      175.902403
2       1.430100     -220.556211
2       1.989300      204.958085
2       2.780900     -130.962041
1       2.842100       45.825781
0      11.049000        5.996205
8                                                        D-G block (AREP)
2        .432300        -.218699
2       1.632300      -29.019674
2       1.959400       86.319069
2       2.657300     -154.261861
2       3.919600      263.326014
2       5.707100     -188.003359
1       9.006300       66.478281
0      45.861500        6.709649
8                                                        F-G block (AREP)
2        .597400       19.797970
2        .679300      -62.511195
2        .860600      116.891258
2       1.185700     -202.106483
2       1.719400      274.155344
2       2.387800     -150.327522
1       3.857300       49.980030
0      13.488900        9.284637
8                                                        block (SO potential)
2        .290400         .011998
2        .920100      -15.887332
2       1.084500       62.388680
2       1.430100      -97.567302
2       1.989300       77.922050
2       2.780900      -34.032450
1       2.842100        -.904480
0      11.049000        -.295330
8                                                        block (SO potential)
2        .432300        -.037005
2       1.632300        4.031599
2       1.959400       -9.421223
2       2.657300        8.472723
2       3.919600        1.512927
2       5.707100       -7.754193
1       9.006300         .432276
0      45.861500        -.060172
8                                                        block (SO potential)
2        .597400        -.154263
2        .679300         .474888
2        .860600        -.825680
2       1.185700        1.122090
2       1.719400        -.989723
2       2.387800         .336488
1       3.857300         .007931
0      13.488900         .000568
6                                                        block (SO potential)
2       1.188500        -.020081
2       2.826400         .052528
2       8.247200        -.104978
2      20.529000         .432037
2      70.763200         .241122
1     223.323400        1.513462
68                                                       Number of core electrons

ECPCE68_DZ        Name of basis set
4                             Highest angular momentum + 1 (3+1 in this case)
4  1  0                     #primitive, orbital angular momentum + 1, #contraction (0 for uncontracted)
 0.5323                   exponent
 0.3023
 0.0964
 0.0370
5  2  0                       P block
 0.7128
 0.2248
 0.0994
 0.0459
 0.0213
5  3  0                        D block
 9.9370
 1.6752
 0.7170
 0.3090
 0.1276
1  4  0                        F block
 0.9126

cc-pVDZ
2
4    1   2                                             Here, contracted basis set is used (#contraction = 2)
13.010     0.019685    0.000000        Additional two columns containing construction coefficients are given
1.9620     0.137977    0.000000
0.4446     0.478148    0.000000
0.1220     0.501240    1.000000
1    2   1
0.7270     1.000000

Sample2: Tl2 (r=3.0 Å)
&ao
 coord =
  “Tl” 0.0 0.0 0.0
  “Tl” 0.0 0.0 3.0
 basis =
  “Tl” “ECPCE68SO_DZ” “ECPCE68SO”
 gaussian = “S”
&end

Sample3: Pb atom with 78-core SOREP (4 valence electrons)
&ao
 coord =
  “Pb” 0.0 0.0 0.0
 basis =
  “Pb” “ECPDS78MDFSO_DZ” “ECPDS78MDFSO”
 gaussian = “S”
&end


Input group: geom

Geometry scan input.

step integer
number of scan
Default: 1

tlist integer(:)
list of target atom for displacement
should match with coord variable
Default: 0

dx real
∆x for each step
Choice: 0 < dx < any real number, 
e.g., dx= 0.1 (0.1Å displacement)
Default: 0.0d0

dy real
∆y, see dx
Default: 0.0d0

dz real
∆z, see dx
Default: 0.0d0

Sample: TlH bond scan (10 points, 0.1Å displacement of H at a time), see Sample2 of ao group for molecular specication
&geom
 step=10
 tlist=2
 dx=0.0
 dy=0.0
 dz=0.1
&end


Input group: scf

Ncharge integer
molecular charge
Choice: e.g., 0 for neutral, 1 for cation, -1 for anion
Default: 0

diis integer
set the iteration number at which DIIS starts
Choice: 0 will disable DIIS acceleration
Default: 3

maxb integer
maximum dimension of b space in DIIS
Default: 6

Felec integer
number of frozen electrons (in development)
Default: 0

mxiter integer
maximum number of iteration
Default: 5000

Ethresh real
Energy threshold for SCF convergence
Default: 1.0d-10 (i.e. 10-10 a.u.)

reordms real(:)
molecular spinor reordering list (in development)
Default: 0

A successful run of KPACK will generate a SCF coefficient file that can be used as an initial guess for a subsequent calculation. This can be achieved by issuing an argument in the command line, e.g.,

#kpack input_file restart >& output_file &

The coefficient file must be present in the same directory as the input.

Sample: +1 ion, with lower threshold
&scf
 Ncharge=1
 Ethresh=1.0d-8
&end


Input group: trans

Hspnr integer
index of highest spinor to be included in post-HF calculation
number of spinors rather than number of Kramers pairs is used
Default: 0 (this will produce error!)

Sample: for post-HF calculation to include 34 Kramers pairs
&trans
 Hspnr=68
&end


Input group: ci

Nexce integer
excitation level
Choice: 1 for CIS, 2 for CISD, 3 for CISD, and so on.
Default: 0 (this will produce error!)

Nroot integer
number of CI roots 
Default: 1

mxiter integer
maximum direct CI iteration
Default: 50

mnsp integer
minimum zeroth-order space in Davidson algorithm
often set equal to Nroot or slighter larger
Default: 1

mxsp integer
maximum zeroth-order space in Davidson algorithm
when the space is full, the space is reconstructed,
this may delay the convergence.
if large memory is available, increase this number
Default: 10

Advanced input
CI_Belec integer
number of electrons in the zeroth-order space

CI_Bact2 integer
size of zeroth-order space

Sample: CISDTQ
&ci
 Nexce=4
 Nroot=1
 mxiter=30
 mxsp=50
&end


Input group: casscf

Nroot integer
number of KRCASSCF roots
Default: 1 

space integer(3)
defines the inactive, active, external spaces
cumulative spinor (not KR) indices are used.
Default: 0

casmxit integer
Maximum KRCASSCF iteration number
Default: 20

cimxit integer
Maximum KRCI iteration (micro iteration) number
Default: 20

mnsp integer
see ci group
Default: 1

mxsp integer
see ci group
Default: 20

Advanced input
CAS_elec integer
override the number of electrons in post-HF stage
useful for open-shell calculation
Default: 0

CAS_root integer(2)
state-averaging parameter
Choice: first root, last root
Default: 0

CAS_expmat integer
determines the construction method for an exponential matrix 
Choice: 0 for Taylor expansion, 1 for Dalgaard method
Default: 0

CAS_cistop logical
stops after initial CASCI calculation
Default: .false.

CAS_manual logical
manual state-averaging by user during the calculation
Default: .false.

CAS_aver integer(:)
override state-averaging scheme 
Default: 0

CAS_extra integer
extra roots calculation for CASSCF/CI (in development)
Default: 0

Sample: SA-KRCASSCF (15roots) of Pb atom (see Sample3 of ao group) and 6p active space
&casscf
 Nroot=15
 space=2,8,48 !1-2:6s(inactive), 3-8:6p(active), 9-48:external
&end