# Estimating diffusion constant for AWH simulation
**URL:** https://gromacs.bioexcel.eu/t/estimating-diffusion-constant-for-awh-simulation/11071
**Category:** User discussions
**Tags:** awh
**Created:** [January 18, 2025, 2:34pm UTC](https://gromacs.bioexcel.eu/t/estimating-diffusion-constant-for-awh-simulation/11071 "2025-01-18T14:34:01Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@roozi](https://gromacs.bioexcel.eu/u/roozi)
#### Post date: [January 18, 2025, 2:34pm UTC](https://gromacs.bioexcel.eu/t/estimating-diffusion-constant-for-awh-simulation/11071/1 "2025-01-18T14:34:01Z")
GROMACS version:2024.3
GROMACS modification: No
Hi
I want to run a protein in water simulation to estimate some PMF using AWH method.
As i know, Diffusion constant (D) is a key .mdp file entry to secure a fast and yet accurate enough simulation. Gromacs reference manual on this subject reads:
> …For good convergence, it should be chosen as large as possible (while maintaining a stable system) giving large initial bias updates and fast initial transitions… It may be a good idea to run a short trial simulation and after the first covering…An accurate estimate of the diffusion can be obtained from an AWH simulation with the [gmx awh] tool.
but it does not elaborate on how to extract the estimated diffusion constant from an earlier AWH simulation data using [gmx awh] command line tool.
i will appreciate if someone can help me on this.
Ps. I already have the .edr & log files of the initial AWH simulation with some rounds of covering and wish to use it to estimate D for a fresh AWH simulation.
kind regards,
roozi
---
### Author:  [@hess](https://gromacs.bioexcel.eu/u/hess)
#### Post date: [January 28, 2025, 11:42am UTC](https://gromacs.bioexcel.eu/t/estimating-diffusion-constant-for-awh-simulation/11071/2 "2025-01-28T11:42:25Z")
The reference manual describes this, formula (360). The diffusion is given by D(lambda)=kT eta^-1(lambda), where eta is the friction metric. The friction metric can be obtained with gmx awh -fric