[gmx-users] (no subject)

[hugo verli]

Hi,

I have obtained an error already described and explained in this mail list. 
However I am using only double precision and the minimization with conjugated 
gradients gives me the same error. Could this error have another reason to 
occur? 

Thanks in advance,

Hugo.


>"......
>Step 1028, E-Pot = -4.2400537500e+05, F-max =  2.46631e+02
>Step 1029, E-Pot = -4.2410303125e+05, F-max =  8.04250e+02
>Negative w:  -4.807500339610e+13
>z=  -7.589670500000e+06
>gpa=   5.727671611336e+06, gpb=   1.845044696754e+07
>a=   0.000000000000e+00, b=   9.266461607997e-07
>EpotA=  -4.241030312500e+05, EpotB=  -4.240932187500e+05
>Negative number for sqrt encountered (-48075003396096.000000)
>Terminating minimization
> 
>writing lowest energy coordinates.
>Maximum force:  1.31781e+03
> 
>Conjugate Gradients did not converge in 5000 steps
>  Function value at minimum =  -4.2410e+05"
>
>>
>>You are probably using conjugate gradients with single precision mdrun.
>>This can be problematic for algorithmic reasons. If you are just preparing
>>for a MD run, conjugate gradients is overkill, and you can use steep.
>>
>>Conjugate gradients minimizes better, but you typically need this only
>>when preparing for normal mode analysis, in which case you need doube
>>precision anyway.