Solving for a large Number of Roots in EOM-CCSD calculation


Clicked A Few Times
Hello,

I am wondering if there are any special considerations to take when running an EOM-CCSD calculation with 25 roots to solve.

My system is relatively small, I have ~ 150 basis functions, and I am running the calculation on a single node with 64 cores and 128 GB of memory. However the calculation fails just before the EOM-CCSD right-hand side iterations begin.

 100a   (alpha) ---    83a   (alpha)        0.1118289896
   100a   (alpha) ---    91a   (alpha)       -0.1361817718
   101a   (alpha) ---    83a   (alpha)       -0.2961984723
   111a   (alpha) ---    69a   (alpha)       -0.1002392376

 INITIAL STARTS -----
0:CreateSharedRegion:kr_malloc failed KB=: 166180
(rank:0 hostname:odin29.rutgers.edu pid:21791):ARMCI DASSERT fail. ../../ga-5-1/armci/src/memory/shmem.c:Create_Shared_Region():1201 cond:0
Last System Error Message from Task 0:: No such file or directory


When I run the calculation requesting fewer roots be solved the calculation terminates normally. I tried changing the memory setting (change the ratio of stack and global), but that didnt help.

I am using NWChem version 6.3. Below is the input file


START myMolecule
memory stack 600 mb heap 100 mb global 1200 mb
geometry units angstroms print xyz autosym
   C        7.57510        1.55960       -0.10280
   C        6.62530        2.13070        0.73810
   C        5.35710        1.58450        0.72750
   C        5.05280        0.52580       -0.07470
   C        5.97390       -0.06280       -0.91810
   C        7.25470        0.47750       -0.92300
   N        4.27110        1.94830        1.43260
   C        3.25530        1.12250        1.08480
   C        3.73820        0.22740        0.14440
   C        2.99690       -0.84570       -0.51290
   O        3.52350       -1.47120       -1.40490
   N        1.70920       -1.14440       -0.13090
   C        1.02120       -2.21830       -0.85590
   C       -0.41540       -2.30320       -0.41920
   O       -1.30170       -2.39360       -1.23890
   N       -0.71280       -2.28850        0.92500
   C       -2.11580       -2.34180        1.35160
   C       -2.87930       -1.19760        0.73630
   O       -2.32730       -0.42160       -0.01070
   N       -4.22240       -1.04780        1.01510
   C       -5.00190       -0.01040        0.45890
   C       -6.34380        0.06980        0.79100
   C       -7.13180        1.07410        0.25630
   C       -6.57790        2.00110       -0.61090
   C       -5.23400        1.92650       -0.94300
   C       -4.44600        0.91970       -0.40830
   O       -7.35640        2.98420       -1.13490
   H        8.60570        1.97410       -0.11930
   H        6.87100        2.99250        1.39490
   H        5.69850       -0.92670       -1.55990
   H        8.02950        0.04040       -1.58870
   H        4.22470        2.73690        2.13050
   H        2.22310        1.17010        1.49300
   H        1.28170       -0.59160        0.65800
   H        1.07120       -2.00340       -1.94480
   H        1.53620       -3.17950       -0.64360
   H        0.09730       -2.23860        1.59740
   H       -2.55400       -3.30890        1.02430
   H       -2.15700       -2.27200        2.45970
   H       -4.66480       -1.74450        1.67070
   H       -6.78900       -0.67270        1.48730
   H       -8.20840        1.13620        0.52320
   H       -4.78930        2.67210       -1.63630
   H       -3.36920        0.85830       -0.67470
   H       -6.95520        3.66360       -1.76870
end
BASIS
 * library sto-3g
END
SCF
 SINGLET
 RHF
END
TCE
 CCSD
 DIPOLE
 FREEZE CORE ATOMIC
 TILESIZE 15
 DIIS 8
 EOMSOL 2
 NROOTS 25
 thresh 1.0d-5
END
set tce:thresheom 1.0d-4
TASK TCE ENERGY

Clicked A Few Times
Hi,
I would need output file to see what is going wrong. Looks like you are experiencing problems with generating initial starts for the EOMCCSD procedure. Please sent output to karol.kowalski@pnnl.gov
Best,
Karol


Forum >> NWChem's corner >> Running NWChem