<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<span>Sorry, the link https://gitlab.com/gromacs/gromacs/-/issues<br>
</span>
<div>Does not work for me. I have created an account at gitlab,<br>
</div>
<div>but it results in PAGE NOT FOUND<br>
</div>
<div><br>
</div>
<div><br>
</div>
<div>>Hello,<br>
</div>
<div>><br>
</div>
<div>>can you please open an issue for this on <br>
</div>
<div>>https://gitlab.com/gromacs/gromacs/-/issues.<br>
</div>
<div>><br>
</div>
<div>>This looks like something that needs immediate attention for the 2020.2 <br>
</div>
<div>>patch release.<br>
</div>
<div>><br>
</div>
<div>>Thanks!<br>
</div>
<div>><br>
</div>
<div>>Paul<br>
</div>
<div>><br>
</div>
<div>>On 06/04/2020 08:59, Igor Leontyev wrote:<br>
</div>
<div>>><br>
</div>
<div>>> Dear gmx developers,<br>
</div>
<div>>><br>
</div>
<div>>> In solvation free energy simulations with gmx 2020.1 I came across the <br>
</div>
<div>>> following issue. vdW component of dG and DHDL values are ~2X <br>
</div>
<div>>> overestimated at the lambda region in which vdw-lambdas are changing. <br>
</div>
<div>>> Additional testing has shown that this issue is related to "vdw-type = <br>
</div>
<div>>> PME."<br>
</div>
<div>>><br>
</div>
<div>>><br>
</div>
<div>>> I have compared (for lambda-state=4) <dVvdw/dl> results of gmx2020 <br>
</div>
<div>>> with vdw-type: PME and CutOff, with the result of old gmx5.1 with <br>
</div>
<div>>> vdw-type = PME. All was run on CPU for 5 nanosecs with PME <br>
</div>
<div>>> electrostatics. It turned out that gmx2020 with vdw-type=CutOff <br>
</div>
<div>>> reproduces (within the accuracy) the <dVvdw/dl> result of gmx5.1 with <br>
</div>
<div>>> vdw-type = PME, which is expected. In the uniform environment both <br>
</div>
<div>>> vdw-types should give accurate solvation FE. The result of gmx2020 <br>
</div>
<div>>> with vdw-type=PME, however, is around 2X exaggerated, which seems to <br>
</div>
<div>>> be a bug.<br>
</div>
<div>>><br>
</div>
<div>>><br>
</div>
<div>>> Here are the numbers:<br>
</div>
<div>>><br>
</div>
<div>>> gmx-ver/vdw-type <dVcoul/dl> <dVvdw/dl><br>
</div>
<div>>><br>
</div>
<div>>> 2020/PME -6.39450e-01 9.72647e+01<br>
</div>
<div>>><br>
</div>
<div>>> 2020/CutOff -3.99985e-01 4.39876e+01<br>
</div>
<div>>><br>
</div>
<div>>> 5.1/PME -3.08675e-01 4.36738e+01<br>
</div>
<div>>><br>
</div>
<div>>><br>
</div>
<div>>> Note, the problem is related to DHDL only, all potential energy <br>
</div>
<div>>> components (including vdW) in all runs are consistent. Let me know if <br>
</div>
<div>>> you need input files to reproduce the issue. Though I believe, it can <br>
</div>
<div>>> be observed for any simple solute in water solution.<br>
</div>
<div>>><br>
</div>
<div>>><br>
</div>
<div>>><br>
</div>
<div>><br>
</div>
<div>>-- <br>
</div>
<div>>Paul Bauer, PhD<br>
</div>
<span></span><br>
</body>
</html>