<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">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<p data-sourcepos="2:1-2:289" dir="auto" class="ContentPasted0">Dear gmx developers,</p>
<p data-sourcepos="2:1-2:289" dir="auto" class="ContentPasted0"><br>
</p>
<p data-sourcepos="2:1-2:289" dir="auto" class="ContentPasted0">Simulation of the Relative Free Energy on GPUs (using gmx 2020-2022.3) gives dramatically different results between BAR and TI (Thermodynamic Integration) methods. The problem is observed in all
tested mutations even in the simplest 1-atom mutation of Acetone->Acetamide in Water with PME:</p>
<p data-sourcepos="2:1-2:289" dir="auto" class="ContentPasted0"><br>
</p>
<p data-sourcepos="4:1-4:58" dir="auto" class="ContentPasted0">GPU results: BAR -173.70 +/- 0.12 TI -181.7074 +/- 0.1501</p>
<p data-sourcepos="6:1-6:58" dir="auto" class="ContentPasted0">cpu results: BAR -181.42 +/- 0.10 TI -181.5248 +/- 0.1512</p>
<p data-sourcepos="6:1-6:58" dir="auto" class="ContentPasted0"><br>
</p>
<p data-sourcepos="8:1-8:119" dir="auto" class="ContentPasted0">As you can see the BAR results disagree between GPU and cpu computations while TI results are in the perfect agreement.</p>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<br>
</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
Sincerely,</div>
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
Igor Leontyev<br>
</div>
</body>
</html>