Re: [AMBER] CUDA INSTALLATION FAILED (Error: Symbol ‘atompositions_fce’ at (1) has no IMPLICIT type)

From: Akash Deep Biswas <akash.biswas.unimi.it>
Date: Fri, 05 Nov 2021 14:17:37 +0100

Hi,

I avoided the installation through run_cmake and followed the old process of configuring the file, compiling them, and finally installing it.

I could successfully install all the three, serial, parallel and cuda versions.

Using the same OS: Ubuntu 21.10 and Compilers: gcc version 11.2.0

Regards
Akash
PS: If you need any more information or testing, I will remain available.

On 04/11/21 20:56, David A Case <david.case.rutgers.edu> wrote:
>
> On Thu, Nov 04, 2021, Akash Deep Biswas wrote:
> >
> >OS: Ubuntu 21.10
> >Fortran Compiler: gcc version 11.2.0
>
> I suspect that you are seeing something with gcc 11.2 that developers have
> not yet tried. Ubuntu 21.10 is using a very new software stack -- that's
> a characteristic of Ubuntu releases. I'm at 21.04, and things work fine
> there.
>
> One could devise a workaround for this particular problem, but I suspect
> other glitches would show up elsewhere. We developers will have to see who
> can spin up a VM with 20.10 and tackle the issues that appear. Volunteers
> are welcome.
>
> ...thx...dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Nov 05 2021 - 06:30:03 PDT
Custom Search