I am trying to build pmemd.cuda.MPI but the error mentioned above occurred.
So where should we mention the CUDA_HOME variable. echo $CUDA_HOME gives no
result. And also do we need any other file to build the mpi version.
On Thu, Jan 23, 2020 at 3:40 PM Sruthi Sudhakar <
sruthisudhakarraji.gmail.com> wrote:
> I have even run some minimization and heating jobs using pmemd.cuda and
> they were completed fine.
>
> On Thu, Jan 23, 2020 at 3:38 PM Sruthi Sudhakar <
> sruthisudhakarraji.gmail.com> wrote:
>
>> the result of which pmemd.cuda is :
>> /apps/amber18/bin/pmemd.cuda
>>
>> On Thu, Jan 23, 2020 at 3:03 PM Elvis Martis <elvis_bcp.elvismartis.in>
>> wrote:
>>
>>> Is the serial version for cuda also installed??
>>> The error says CUDA_HOME is not set and the serial cuda installation will
>>> fail because of the same error.
>>> What is the output you get after typing “which pmemd.cuda” ? ( without
>>> quotes)
>>> On Thursday, January 23, 2020, Sruthi Sudhakar <
>>> sruthisudhakarraji.gmail.com>
>>> wrote:
>>>
>>> > But Serial version is already installed.
>>> >
>>> > On Thu, 23 Jan 2020 at 1:36 PM, Elvis Martis <elvis_bcp.elvismartis.in
>>> >
>>> > wrote:
>>> >
>>> > > You will have to define CUDA_HOME as the error says. Install CUDA
>>> serial
>>> > > first
>>> > > ./configure -cuda gnu
>>> > > On Thursday, January 23, 2020, Sruthi Sudhakar <
>>> > > sruthisudhakarraji.gmail.com>
>>> > > wrote:
>>> > >
>>> > > > Respected group members,
>>> > > >
>>> > > > Previously I have written a mail regarding the installation of GPU
>>> > > > accelerated version of AMBER and was suggested to first do the
>>> single
>>> > gpu
>>> > > > installation at first. That installation was successful and jobs
>>> are
>>> > > > running smoothly using pmemd.cuda. But now we want to do the
>>> parallel
>>> > > > installation as well. Should we begin the installation from the
>>> > beginning
>>> > > > or just follow the next step from the manual, that is begin from
>>> > > >
>>> > > > ./configure -cuda -mpi -gnu
>>> > > > When I continued with this step , I got the following error
>>> message:
>>> > > > Error: CUDA_HOME is not set. This must point to your NVIDIA tools
>>> > > > installation
>>> > > > Configure failed due to the errors above!
>>> > > > Please suggest how to proceed further.
>>> > > >
>>> > > >
>>> > > > Thank you
>>> > > > _______________________________________________
>>> > > > AMBER mailing list
>>> > > > AMBER.ambermd.org
>>> > > > http://lists.ambermd.org/mailman/listinfo/amber
>>> > > >
>>> > >
>>> > >
>>> > > --
>>> > > Best Regards
>>> > > Elvis Martis
>>> > > Mumbai.
>>> > > _______________________________________________
>>> > > 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
>>> >
>>>
>>>
>>> --
>>> Best Regards
>>> Elvis Martis
>>> Mumbai.
>>> _______________________________________________
>>> 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 Thu Jan 23 2020 - 03:30:01 PST
