[gmx-users] Temperature increase

[Mark Abraham]

Zuzana Benkova wrote:
> I have run an NVE simulation of decamer of glycin first with PBC 
> (coulombtype=Cut-off) and later without PBC (coulombtype=PME).

You can't have a non-periodic system with PME.

> The 
> gen-temp was set to 300K with option gen-vel=yes. The temperature start 
> to raise during simulation and finaly the following messages appeared in 
> log file
> 
> Shake did not converge in 1000 steps
> 
>     i     mi      j     mj      before       after   should be
> 
>    66  14.01     67   1.01     0.10100     0.10318     0.10100
> 
> Constraint error in algorithm Shake at step 980454
> 
> Wrote pdb files with previous and current coordinates
> 
> How can I cope with this?

We can't tell, because the above description is not self-consistent.

In any case you haven't defined your topology, or minimized your 
structure or equilibrated properly. Then some bad structural element is 
causing a region to have too much energy which is breaking structures. 
Search the mailing list archives for suggested equilibration regimes, or 
read relevant literature examples, or do tutorial material.

Mark