GRRM is a code to perform a local, semiglobal, or global reaction path exploration. It is very useful for many purposes from a conventional reaction path calculation to an exhaustive reaction path exploration and a complex rection path network construction. It has been applied to various reaction systems such as organic reaction, organometallic catalysis, cluster catalysis, radical reaction, photoreaction involving electronic excited states, crystal phase transition under periodic boundary conditions, enzyme catalysis described by QM/MM-ONIOM method, and so on. GRRM includes internal interfaces with Gaussian03/09/16, molpro, GAMESS, ORCA, TURBOMOLE, and SIESTA, and can also be combined with any electronic structure calculation code by preparing a simple code. There are many new features in the version GRRM20, and four major ones are described below.

Update information

25th Feburary, 2022

  • ▸ Fixed a bug that occurred in the “11th January, 2022” version when introducing the function to reduce the traffic between nodes in RePATH calculation.

11th January, 2022

  • The following changes have been made:
  • ▸ Optimized to reduce the exchange of data between nodes. Therefore, the speed of massively parallel computing is improved. In particular, in massively parallel RePATH calculation, the obstacle that the calculation is stagnant due to the exchange of data between nodes has been removed.
  • ▸ In this release, the LUP calculation is performed in order of longest route in the RePATH calculation. This eliminates the problem that the calculation for only a small number of routes observed at the end of the massively parallel RePATH calculation continues for a long time.
  • ▸ Fixed a bug caused by the input file name when using DontKeepGauCHK.
  • ▸ In this release, SC=CheckReflect is turned ON by default. If you want to turn it off, add SC=DontCheckReflect.

28th December, 2021

24th December, 2021


Commercial GRRM with a number of new features

See the following references for new features:
S. Maeda, Y. Harabuchi, Exploring paths of chemical transformations in molecular and periodic systems: An approach utilizing force., WIREs Comput. Mol. Sci., 2021, 11, e1538 (23 pages).

You can see all of the chemical reactions.

With the input of just one molecule, GRRM can automatically explore for all the products that can be produced from them, all the reactants that can produce them, and all their reaction pathways. It can be used in various fields such as catalyst design and material screening.

Number of stable structures
(Reactant / Product / Intermediate)
Number of elementary reactions
acetic acid
propionic acid
methyl nitrate


Kinetic simulation / kinetics-based navigation using RCMC

A kinetic analysis method, rate constant matrix contraction (RCMC), which is applicable to complex reaction path networks is available in GRRM20. RCMC can be used in two purposes: (1) as a kinetic navigation of the SC-AFIR search, and (2) analysis of a reaction path network obtained by a SC-AFIR search. The kinetic navigation enables to perform an on-the-fly kinetic simulation of a given chemical system under an experimental condition (reaction temperature and reaction time). The latter allows to coarse-grain a reaction path network, to compute an overall rate constant between the reactant and the product, to extract the kinetically most feasible path from a reaction path network, and so on.

Periodic boundary conditions (PBCs)

Geometry optimization, reaction path calculation, and automated reaction path exploration under PBCs are possible in GRRM20. When translational vectors are specified as active, both translational vectors and atom positions are optimized simultaneously. In addition, an automated search for paths of reactions on a 2D/1D slab model can also be performed by fixing all or some of translational vectors.

An optimizer for systems including >500 active atoms

An optimization algorithm for systems including a large number of active variables is available in GRRM20. The algorithm performs a geometry optimization in the full dimension using a limited dimensional PES expressed by gradient vectors obtained during the optimization. Its performance was tested with systems including up to 1000 atoms in combination with a semi-empirical quantum chemical calculation method.

Massive parallelization

In GRRM20, parallelization efficiency of SC-AFIR calculations was improved significantly. Various jobs that execute 100-500 path calculations simultaneously have been done in the developer group. In these jobs, several cores have been used in each path calculation, and thus the total numbers of cores used in these jobs have been 1000~2000.


Many fine optimizations and speedups have been made, backed by the state-of-the-art practical achievements of GRRM developers.

Interface for changing AFIR search method from the outside

With a simple external script, you can use the option to reflect informatics methods and empirical rules in the search procedure, and users can also participate in the development of methods for accelerating automatic search.

Applicable for various R&D

For example, drug synthesis. By clarifying all reaction pathways that can synthesize the desired drug molecule, it can be used for catalyst design for rate-determining reactions and suppression of by-products.

For example, designing luminescent materials. Since the stability of the material can be considered from the viewpoint of reaction, it can be used for designing long-life materials that emit stable light.

For example, a combustion reaction. Thousands and tens of thousands of elementary reactions can be obtained based on highly accurate quantum chemistry calculations, so it is possible to meet the accuracy of reaction speed in car and rocket engine design CAE.

Price, Specification and Support


You can purchase an annual (per year) license. For more information, please contact us using this form. When making inquiries, please enter the email address of your organization, not the free email address.


Operating environment #1- Hardware -
x86_64 computer that works below

- OS -
Red Hat Enterprise Linux 7.x or CentOS 7.x
Red Hat Enterprise Linux 8.x, CentOS 8.x or AlmaLinux 8.x
(Red Hat Enterprise Linux 6.x and CentOS 6.x are not supported.)

- Required software -
Gaussian16 or Gaussian09 *1

- Optional software -
Gaussian03, Molpro, GAMESS, ORCA, Turbomole, SIESTA *2 *3
Operating environment #2- Hardware -
Computation nodes of the Supercomputer "Fugaku"

- Required software -
Gaussian16 *4
Included itemsSoftware *5

*1 Not included in this product. Please prepare in advance.
*2 Not included in this product. If you want to use it, please prepare it in advance.
*3 GRRM20 has the general interface with external ab initio programs.
*4 GRRM20 and Gaussian16 on the Supercomputer “Fugaku” are available through our Science Cloud service, specifically, “Fugaku” type of its “standard” plan.
*5 Please refer to the Manual page on the AFIR site for the latest GRRM20 Users Manual.


We support you for troubles in setting up and starting GRRM20 for free of charge.

For how to use GRRM20 and how to interpret the results, please contact the Forum on the AFIR site by yourself. (Please note: when purchasing the GRRM20 license, please register as a user on the AFIR site).

Please refer to the Manual page on the AFIR site for the latest GRRM20 Users Manual.

GRRM program function comparison table

GRRM program function comparison table is here (PDF).

GRRM is a registered trademark of Institute for Quantum Chemical Exploration .

Contact us

  • キャンペーン情報
  • ご購入前のお問合せ
  • 見積り依頼



平日9:30~17:30 (土曜日、日曜日、祝祭日、年末年始、夏期休暇は、休日とさせていただきます。)