TY - JOUR
T1 - Beyond Walkers in Stochastic Quantum Chemistry
T2 - Reducing Error Using Fast Randomized Iteration
AU - Greene, Samuel M.
AU - Webber, Robert J.
AU - Weare, Jonathan
AU - Berkelbach, Timothy C.
N1 - Funding Information:
We thank Aaron Dinner and Sandeep Sharma for useful discussions about this work, and we thank Anthony Scemama for noting an error in the geometry of H 2 O used in calculations in a preprint. S.M.G. and T.C.B. were supported by start-up funding from the University of Chicago and by the Flatiron Institute. The Flatiron Institute is a division of the Simons Foundation. R.J.W. and J.W. were supported by the Advanced Scientific Computing Research program through award DE-SC0014205. R.J.W. was also supported by NSF RTG award 1547396 at the University of Chicago and by a MacCracken Fellowship and NSF RTG award 1646339 at New York University. Calculations were performed with resources provided by the University of Chicago Research Computing Center.
Publisher Copyright:
Copyright © 2019 American Chemical Society.
PY - 2019/9/10
Y1 - 2019/9/10
N2 - We introduce a family of methods for the full configuration interaction problem in quantum chemistry, based on the fast randomized iteration (FRI) framework [Lim, L.-H.; Weare, J. SIAM Rev. 2017, 59, 547; DOI: 10.1137/15M1040827]. These methods, which we term "FCI-FRI", stochastically impose sparsity during iterations of the power method and can be viewed as a generalization of full configuration interaction quantum Monte Carlo (FCIQMC) without walkers. In addition to the multinomial scheme commonly used to sample excitations in FCIQMC, we present a systematic scheme where excitations are not sampled independently. Performing ground-state calculations on five small molecules at fixed cost, we find that the systematic FCI-FRI scheme is 11-45 times more statistically efficient than the multinomial FCI-FRI scheme, which is in turn 1.4-178 times more statistically efficient than the original FCIQMC algorithm.
AB - We introduce a family of methods for the full configuration interaction problem in quantum chemistry, based on the fast randomized iteration (FRI) framework [Lim, L.-H.; Weare, J. SIAM Rev. 2017, 59, 547; DOI: 10.1137/15M1040827]. These methods, which we term "FCI-FRI", stochastically impose sparsity during iterations of the power method and can be viewed as a generalization of full configuration interaction quantum Monte Carlo (FCIQMC) without walkers. In addition to the multinomial scheme commonly used to sample excitations in FCIQMC, we present a systematic scheme where excitations are not sampled independently. Performing ground-state calculations on five small molecules at fixed cost, we find that the systematic FCI-FRI scheme is 11-45 times more statistically efficient than the multinomial FCI-FRI scheme, which is in turn 1.4-178 times more statistically efficient than the original FCIQMC algorithm.
UR - http://www.scopus.com/inward/record.url?scp=85071998437&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85071998437&partnerID=8YFLogxK
U2 - 10.1021/acs.jctc.9b00422
DO - 10.1021/acs.jctc.9b00422
M3 - Article
C2 - 31390198
AN - SCOPUS:85071998437
SN - 1549-9618
VL - 15
SP - 4834
EP - 4850
JO - Journal of chemical theory and computation
JF - Journal of chemical theory and computation
IS - 9
ER -