-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathREADME
83 lines (53 loc) · 2.42 KB
/
README
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
=== Running ParaGauss ===
For parallel calculation in batch mode use the Scheme script
./runqm
derived from ./guile/runqm.scm by inserting specific search path
during build process. This script does its best to emulate the
behaviour of the traditional "runpg" bash script. Use it as in
mpirun -np 8 $PWD/runqm [INPUTS...]
Note that neither "runqm" script nor "guile-qm" interpreter are in
your $PATH. So that you will need to specify the path in full
occasionally. In case of troubles, check the "shebang" line (first
line starting with #!) at the top of "runqm" and the "%load-path" a
few lines below.
There is a possibility to run ParaGauss from a Scheme interpreter.
The default (WITH_GUILE=1) build produces an executable "guile-qm" in
the top directory (where this README file resides) that can be used as
a standalone Scheme interpreter. You may try to execute it in serial:
$ ./guile-qm -L ./
guile> (+ 1 2)
3
and run PG from within it:
guile> (use-modules (guile paragauss))
guile> (define world (qm-init))
guile> (copy-file "some-input" "input")
guile> (qm-run world)
guile> (qm-finalize world)
Or write your own interpreted scripts ...
=== Fallback operation mode ===
If you have troubles with Scheme, set WITH_GUILE=0 in the Makefile and
recompile. Start ParaGauss serially as in
./mainscf_VX.Y.Z
The program expects the input file named
input
to reside in the current directory.
Beware there will be a lot of temporary files created in the current
directory together with the output files. Specify location for the
temporary files by setting environment variable
export TTFSTMP=/tmp
or in the command line
TTFSTMP=/tmp ./mainscf_VX.Y.Z
For MPI-enabled builds use mpirun/mpiexec as appropriate:
TTFSTMP=/tmp mpirun -np 8 $PWD/mainscf_VX.Y.Z
You can specify output directory and the alternative name of the input
file as in
TTFSTMP=/tmp TTFSOUPUTDIR=$PWD/o.test TTFSINPUT=i.test ./mainscf_VX.Y.Z
Note that the directory, if specified, must already exist. Moreover
it must exist on all nodes potentially involved in a parallel
computation.
The startup script located in
./bin/runpg
attempts to guess execution nodes and create directories prior to
invoking ParaGauss. However, the startup procedure varies from one
setup to another so that the startup script may or may not work on
your platform.