Output files

Recent site activity

Titan at OLCF
edited by Jeongnim Kim
Benchmarks on HPC
edited by Jeongnim Kim
Titan at OLCF
edited by Jeongnim Kim
Cray XT
edited by Jeongnim Kim
Building and tuning einspline
attachment from Jeongnim Kim
View All

For each <qmc> section, files with these extensions will be generated

File Extension	File Type	Contents	Tools
scalar.dat	ASCII	Block-averaged properties Header starting with comment (#)	utils/Energy.py, dataspork(pp) gnuplot, xmgrace
dmc.dat	ASCII	Population properties per step Header starting with comment (#)	Same as scalar.dat
config.h5	HDF5	State of a population for restart including random numbers	hdf tools, e.g., h5dump
stat.h5	HDF5	Collectables	Same as config.h5
storeConfig.h5	HDF5	Data for forward walking	Same as config.h5
cont.xml	XML	Input xml	Any xml utilities

and last, *.cont.xml to assist restarts for the last section.

Analysis Tools

In utils directory, python scripts are available for the analysis of *.dat files.

Need numpy

*.dat files are space-separated table format which can be plotted with gnuplot or xmgrace

Energies are in Hartree
Each line has the block averages for scalar.dat and the MC step averages for dmc.dat
In QMCPACK, a MC step denotes a sweep over all the walkers in a population regardless of the update (walker-by-walker or particle-by-particle) mode.
Always discard the data during a warmup step as shown in the example below.

Calculating the properties using Energy.py

sample.scalar.dat

Energy.py sample.scalar.dat 10

LocalEnergy          = -551.1642 +/- 0.0010
LocalPotential       = -991.655 +/- 0.015
Kinetic              = 440.491 +/- 0.015
ElecElec             = 116.9320 +/- 0.0088
IonIon               = -242.1943087 +/- 0.0000020
LocalECP             = -870.638 +/- 0.023
NonLocalECP          = 4.2448 +/- 0.0033
MPC                  = 117.0465 +/- 0.0088
KEcorr               = 0.01247768 +/- 0.00000000
BlockWeight          = 12027073 +/- 23187
BlockCPU             = 95.97 +/- 0.31
AcceptRatio          = 0.9934108 +/- 0.0000029
-----------------------------------------------------
Corrected energy     = -551.0372 +/- 0.0010

This particular data is for a bulk system and estimators to calculate the finite-size corrections are used during a simulation.

Energy (column 2)	Potential Energy (column 3)	Kinetic Energy (column 4)

sample.dmc.dat

Energy.py sample.dmc.dat 1000

LocalEnergy          = -551.16413 +/- 0.00100
Variance             = 19.6290 +/- 0.0043
Weight               = 120272 +/- 231
NumOfWalkers         = 120272 +/- 231
TrialEnergy          = -551.1643 +/- 0.0016
DiffEff              = 0.98852257 +/- 0.00000050

Note that the first 1000 steps are discarded.