[molpro-user] Failure in CCSD(T) with Molpro 2015
王黎明
wanglm at scut.edu.cn
Mon Jan 18 15:24:58 CET 2016
Dear Molpro User,
We have recently encountered problems with Molpro 2015. I have tested the three binary versions available. Our system is a standalone workstation, with two Intel 2680 V3 CPU (12 core each), 256 GB RAM and 3 TB of HD, and is running CentoOS 6.5. The problem is quite stranger.
On our system, we can successfully finish CCSD(T)-F12 calculations for small system like C10H7Cl (1-chloronaphthalene). But we could never finish CCSD(T) for a terpene C10H16 and other species with similar size. This is our input file:
***,3-Carene, CCSD(T)-F12/cc-pVDZ-F12
MEMORY,2048,MW ! Memory per process
ANGSTROM
GEOMTYP=xyz
geometry={
26 ! Number of Atoms
The Geometry of 3-Carene @ M06-2X/6-311++G(2df,2p)
C, -1.8219980965, 0.2582696038, 1.2531734283
C, -1.2133421422, -0.7693651824, 1.8333515558
C, -1.1914136357, 1.0721687178, 0.1553394121
C, -3.1986687528, 0.7027697572, 1.6532770721
C, 0.1623697877, -1.2717240158, 1.5042768978
C, 0.1766443043, 0.6203095468, -0.3049984511
C, 0.8550741078, -0.5482548519, 0.3693504023
C, 1.4063291417, 0.8424854962, 0.5324569045
C, 2.6868377182, 1.1612611389, -0.2038207099
C, 1.2852109884, 1.5432904839, 1.8645363495
H, -1.7353032617, -1.3050826125, 2.6202749227
H, -1.8740386021, 1.059118132, -0.6993372773
H, -1.1585455093, 2.120004797, 0.4798590559
H, -3.1836186301, 1.7406867039, 1.9951985686
H, -3.6050967798, 0.0842212228, 2.4514159251
H, -3.8817771935, 0.6611496866, 0.8016163978
H, 0.0872204053, -2.3321779251, 1.2509342445
H, 0.779426763, -1.2332303691, 2.4097277505
H, 0.3339887657, 0.7017023158, -1.3729394263
H, 1.4349473754, -1.1911514523, -0.280392755
H, 2.7596488249, 2.2308281512, -0.4135722873
H, 2.7373771196, 0.6281303081, -1.1536820371
H, 3.5582629942, 0.8764150774, 0.389840421
H, 2.122025192, 1.2729259462, 2.5122809229
H, 0.3627043637, 1.2934154641, 2.3855346591
H, 1.3140507518, 2.6265598597, 1.7269300535
}
BASIS=cc-pVDZ-F12
{RHF;}
{RCCSD(T)-F12 SCALE_TRIP=1; CORE,10;}
The calculations halted in several ways:
(1) Sometimes we have，
.............................
Starting RCCSD calculation
ITER. SQ.NORM CORR.ENERGY TOTAL ENERGY ENERGY CHANGE DEN1 VAR(S) VAR(P) DIIS TIME
1 1.52605156 -1.56784118 -389.64205619 -1.56784118 -0.04197524 0.59D-02 0.93D-02 1 1 3159.65
2 1.56713760 -1.60940384 -389.68361885 -0.04156266 -0.00273179 0.15D-03 0.92D-03 2 2 5944.14
3 1.58004829 -1.61713880 -389.69135382 -0.00773496 -0.00023873 0.44D-04 0.76D-04 3 3 7720.22
4 NaN NaN NaN NaN -0.00001972 0.29D-05 0.76D-05 4 4 9379.01
5 NaN NaN NaN NaN NaN NaN NaN 5 5 11009.04
(2) Sometimes we have
.................
ITER. SQ.NORM CORR.ENERGY TOTAL ENERGY ENERGY CHANGE DEN1 VAR(S) VAR(P) DIIS TIME
1 1.52605156 -1.56784118 -389.64205619 -1.56784118 -0.04197524 0.59D-02 0.93D-02 1 1 2926.21
2 1.56713760 -1.60940384 -389.68361885 -0.04156266 -0.00273179 0.15D-03 0.92D-03 2 2 4733.93
3 1.58004829 -1.61713880 -389.69135382 -0.00773496 -0.00023873 0.44D-04 0.76D-04 3 3 6614.77
4 1.58418449 -1.61909890 -389.69331392 -0.00196010 -0.00001972 0.29D-05 0.76D-05 4 4 8559.60
5 1.58494733 -1.61921990 -389.69343491 -0.00012099 -0.00000231 0.72D-06 0.82D-06 5 5 10153.03
6 1.58514342 -1.61924028 -389.69345529 -0.00002038 -0.00000027 0.88D-07 0.95D-07 6 6 11626.68
and then molpro keeps running for another day without any further output. I have to kill the process.
(3) Sometimes we can finish CCSD, and then running for (T)-calculations for days with no further output. I have to kill the process again.
............
Norm of t2 vector: 0.75760045 P-energy: -1.61924958
Alpha-Beta: -1.27028351
Alpha-Alpha: -0.17448303
Beta-Beta: -0.17448303
We have tried various combination to run the calculation, as
molpro -n 12 -G 10000 x.inp &
or molpro -n 12 -G 100000 x.inp &
......
and we have also tested different disks as scratch space to make sure it is not due to hardware problem. We have also carefully monitored the temperatures of hardwares and found no problem there. Now my brain is completely exhausted, and I am asking for your help here. Thanks in advance!
Yours
Liming
--
Liming Wang, Ph D
School of Chemistry & Chemical Engineering
South China University of Technology
381, Wushan Rd.
Guangzhou, China 510640
E-Mail: wanglm at scut.edu.cn
ResearchGate: https://www.researchgate.net/profile/Liming_Wang6
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20160118/0858d3ea/attachment-0001.html>
More information about the Molpro-user
mailing list