Re: Build for CUDA support

From: Renfro, Michael (Renfro_at_tntech.edu)
Date: Sat Mar 03 2018 - 10:03:26 CST

Assuming we’re looking at http://www.ks.uiuc.edu/Research/namd/2.12/notes.html , the section "CUDA GPU Acceleration” includes:

  The CUDA (NVIDIA's graphics processor programming platform) code in
  NAMD is completely self-contained and does not use any of the CUDA
  support features in Charm++. When building NAMD with CUDA support
  you should use the same Charm++ you would use for a non-CUDA build.
  Do NOT add the cuda option to the Charm++ build command line. The
  only changes to the build process needed are to add --with-cuda and
  possibly --cuda-prefix ... to the NAMD config command line.

-- 
Mike Renfro  / HPC Systems Administrator, Information Technology Services
931 372-3601 / Tennessee Tech University
> On Mar 3, 2018, at 8:22 AM, Mahmood Naderan <mahmood.nt_at_gmail.com> wrote:
> 
> Hi,
> I followed the installation guide and successfully installed namd. The document is large so, maybe the question I am asking is explained in the document. 
> The build procedure has nothing to do with the cuda toolkit. Am I right? I followed these steps:
> 
> Getting the source package
> tar zxf NAMD_2.12_Source.tar.gz 
> cd NAMD_2.12_Source/
> tar xf charm-6.7.1.tar
> cd charm-6.7.1
> ./build charm++ multicore-linux64 --with-production
> cd multicore-linux64/tests/charm++/megatest
> make pgm
> ./pgm +p4
> cd ../../../../..
> ./config Linux-x86_64-g++ --charm-arch multicore-linux64
> cd Linux-x86_64-g++
> make -j4
> 
> 
> 
> Regards,
> Mahmood
> 
> 

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:53 CST