mirror of https://github.com/QMCPACK/qmcpack.git
d095727a58
find -name "CMakeLists.txt" -exec cmake-format -i {} \;
find -name "*.cmake" -exec cmake-format -i {} \;
git checkout external_codes
|
||
|---|---|---|
| .. | ||
| qmc-ref | ||
| CMakeLists.txt | ||
| README | ||
| qmc_long_chol.in.xml | ||
| qmc_long_phmsd.in.xml | ||
| qmc_long_thc.in.xml | ||
| qmc_long_uhf.in.xml | ||
| qmc_short_chol.in.xml | ||
| qmc_short_phmsd.in.xml | ||
| qmc_short_thc.in.xml | ||
| qmc_short_uhf.in.xml | ||
README
These tests include short and long AFQMC runs for a two atom diamond
primitive cell at the gamma point. Tests exist for both the Cholesky
and ISDF/THC factorization of the integrals. For the Cholesky
factorization there are tests for RHF, UHF and particle-hole
multi-Slater determinant (PHMSD) trial wavefunctions. For the ISDF
factorization a RHF trial wavefunction is used. The integrals and
trial wavefunctions were generated using pyscf. The relevant scripts
and files are available in tests/afqmc as these files are re-used for
afqmc unit tests.
Reference data from the high block count AFQMC runs is found below:
EnergyEstim__num_real
afqmc cholesky rhf -10.55201 +/- 0.00033
afqmc cholesky uhf -10.55151 +/- 0.00031
afqmc cholesky phmsd -10.55085 +/- 0.00026
afqmc isdf rhf -10.33408 +/- 0.00043
These values can also be found in qmc-ref/qmc_ref.qmca. The scalar.dat
files were too large to include with qmcpack. The first 10 blocks were
discarded from the reference scalar.dat files during analysis with the
qmca script.
Test input files were created by reducing the number of steps from the
included reference by a factor of 10 for the "long" tests and by a
further factor of 10 for the "short" tests.
A reference value for the mean local energy was taken from the high
block count reference run. Expected error bars for tests using
single determinant trial wavefunctions were derived from the high
block count data by multiplying the reference error bars by a factor
of sqrt(10+1) and sqrt(100+1) for the long and short tests,
respectively. The factor of +1 accounts for the error bar intrinsic
to the high block count runs. Similarly for multi determinant case,
the reference error bars were multiplied by a factor of sqrt(20+1) and
sqrt(201+1) for the long and short tests. The tests were made half as
long due to the additional overhead associated with running
multi-determinant calculations.
The target means and error bars for long and short tests are:
refmean referr longerr shorterr
afqmc cholesky rhf -10.55201 +/- 0.00033 0.00109 0.00332
afqmc cholesky uhf -10.55151 +/- 0.00031 0.00103 0.00312
afqmc cholesky phmsd -10.55085 +/- 0.00026 0.00119 0.00369
afqmc isdf rhf -10.33408 +/- 0.00043 0.00143 0.00432