Re: namd-1: Simulation Terminating Early

From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Mon Feb 09 2015 - 13:33:09 CST

>From off-list communication, this is NAMD 2.10b1 for Windows-64 Multicore,
which was affected by an MSVC compiler bug:

  MSVC compiler bug breaks SETTLE implementation on Windows, causing
simulation instability when "rigidBonds water" or "rigidBonds all" is
enabled. Workaround for 2.10b1 is "useSettle no". Fixed in CVS by adding
-D__SSE2__ to Win32 and Win64 compiler flags to enable vector code that
was not being used previously and does not encounter this bug.

The 2.10 final release should work.

Jim

On Sat, 7 Feb 2015, Akshata Rudrapatna wrote:

> Hello NAMD Community,
>
> I posted earlier about not being able to run the example simulation given
> in the NAMD tutorial on the website. I was able to get the simulation
> started and running, but it runs approximately one minute before
> terminating.
>
> The last message in the log is as follows:
>
> REINITIALIZING VELOCITIES AT STEP 100 TO 310 KELVIN.
> TCL: Running for 100 steps
> PRESSURE: 100 -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND
> -1.#IND -1.#IND
> GPRESSURE: 100 -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND -1.#IND
> -1.#IND -1.#IND
> ENERGY: 100 91.8776 243.3225 250.4144 8.5143
> -14084.3172 1235.7567 0.0000 0.0000
> -1.#IND -1.#IND -1.#IND -12254.4316 -1.#IND
> -1.#IND -1.#IND -1.#IND 86856.0000
> -1.#IND -1.#IND
>
> I have tried changing the timestep and the pairlistdist values in the file,
> as well as the length of time the simulation runs for.
>
> Does anyone have any suggestions or tips as to how to resolve this problem?
> I appreciate your time and consideration.
>
> Thanks,
>
> Akshata Rudrapatna
>

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:39 CST