Dear gmx-users, <br><br>Recently I installed Gromacs to a new cluster, SUN X4100, Dual Core AMD Opteron (tm) Processor. <br>However during the test step, errors occur. <br>The following is the entire output of the "<b>$./gmxtest.pl -double all</b>" command <br>
<br> All 16 simple tests PASSED<br> All 14 complex tests PASSED<br> All 63 kernel tests PASSED<br> N Reference This test<br> 10 -39.0984 -32.5043<br> 11 -39.0984 -32.5043<br> 12 -93.8123 -75.8897<br>
13 -93.8123 -75.8897<br> 21 -11007.5 -9915.17<br> There were 5 differences in final energy with the reference file<br> All 45 pdb2gmx tests PASSED<br> pdb2gmx tests <b>FAILED</b><br><br>I guess this seems to be a comparison between the precalculated value(Reference) and the value calculated in my machine(This test). <br>
<br>That what is the probable causes of this problem? <br>I compiled the program again twice, with <i>gnu compilers</i>, gcc.<br>Also the double precision fftw3 library is used.<br>Only the installing directory and the fftw3 library directories are given during configuration. <br>
<br>Additionally, if I use 'xleap' of Amber package and using 'amb2gmx.pl' script instead of 'pdb2gmx' program, is it OK?<br>(By doing this I use Amber force field in Gromacs.)<br>Anyhow, the programs seems to work without out making any troubles, but I don't know whether I am wasting CPU time or not. ^^ <br>
<br>Thank you ~ <br><br>-- <br>--------------------------------------------------<br>'God used beautiful mathematics in creating the world.'<br>-Paul Dirac <br>'But he created too many objects.'<br>-Seungpyo Hong <br>
<br>Seungpyo Hong <br>Master student at PBIL(Protein BioInformatics Lab.) in KAIST, Daejeon Korea<br>Tel. (82)-18-372-2468<br><a href="mailto:sphong_@kaist.ac.kr">sphong_@kaist.ac.kr</a><br><a href="mailto:sp1020@gmail.com">sp1020@gmail.com</a>