Re: Build NAMD with CUDA for unix workstation (centos 7)

From: Daipayan Sarkar (sdaipayan_at_gmail.com)
Date: Wed Sep 20 2017 - 03:04:32 CDT

Thanks for the steps.

revise "CHARM = xxxxxxxxxxx" in Make.config to a absolute pathway

Do I add the path similar to the cuda-prefix?

Regards,

Daipayan

On Wed, Sep 20, 2017 at 2:22 AM, Haohao Fu <fhh2626_at_gmail.com> wrote:

> Suppose you already have fftw and tcl folder in the NAMD directory
>
> Then you can do:
> (build charm)
> tar xvf charm-xxxxx.tar
> cd charm-xxxxx
> ./build
>
> (build NAMD with CUDA)
> cd ..
> ./config Linux-x86_64-g++ --with-cuda --cuda-prefix /xxxxxxxxxxxxxx
> cd Linux-x86_64-g++
> revise "CHARM = xxxxxxxxxxx" in Make.config to a absolute pathway
> make -j
>
> Haohao
>
> 2017-09-20 14:54 GMT+08:00 Daipayan Sarkar <sdaipayan_at_gmail.com>:
>
>> Hi,
>>
>> Can someone share the steps to build a NAMD executable with CUDA enabled
>> for a multicore Unix workstation? The CUDA toolkit, I am using is
>> 'cuda-repo-rhel7-8-0-local-ga2-8.0.61-1.x86_64.rpm' for centos 7 and
>> x86_64.
>>
>> Thanks,
>> Daipayan
>>
>>
>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:36 CST