Re: [AMBER] An error during NPT/production (GPU run)

From: David A Case <>
Date: Sun, 29 Nov 2020 13:39:39 -0500

On Sat, Nov 28, 2020, Женя Елизарова wrote:

>I see. I have sent you files. This letter awaits moderator approval because
>the message body is too big.

You may need to try to create a smaller system to see what fails, or
consider other alternatives to sharing large files.

>Our system consists of 63420 atoms.

I probably need to apoligize here: I read your previous email as
indicating that you thought your system was too small. But since that
is not the case, other suggestions I made are not relevant.

>I also would like to ask some more questions. Which performance you suppose
>sould be for such size of the system on GPU? (for example in ns/day).

You can look at the GPU benchmarks on our web site, and roughtly
interpolate between results for systems that are smaller and those that
are larger.

>Another question is about calculations in a vacuum. Is it possible on GPU?

Look at igb=6 in the manual. But first, be sure that this is really
what you want to do. The Amber programs and force fields are not
designd for vacuum simulations. So be sure that you have a good reason
for doing this.


AMBER mailing list
Received on Sun Nov 29 2020 - 11:00:03 PST
Custom Search