CUDA

by

Turning off and on ECC RAM for NVIDIA GP-GPU Cards

From NVIDIA Developer site.

Turn off ECC (C2050 and later). ECC can cost you up to 10% in performance and hurts parallel scaling. You should verify that your GPUs are working correctly, and not giving ECC errors for example before attempting this. You can turn this off on Fermi based cards and later by running the following command for each GPU ID as root, followed by a reboot:

Extensive testing of AMBER on a wide range of hardware has established that ECC has little to no benefit on the reliability of AMBER simulations. This is part of the reason it is acceptable (see recommended hardware) to use the GeForce gaming cards for AMBER simulations.

To Turn off the ECC RAM, just do a

# nvidia-smi -g 0 --ecc-config=0
(repeat with -g x for each GPU ID)

To Turn back on ECC RAM, just do

# nvidia-smi -g 0 --ecc-config=1
(repeat with -g x for each GPU ID)
by

Compiling OpenMPI 1.7.2 with CUDA and Intel Compilers 13

If you are intending to compile OpenMPI with CUDA Support, do note that you have to download the feature version of OpenMPI. The version I used for compiling OpenMPI with CUDA is version 1.7.2. The current stable version of OpenMPI 1.6.5 does not have CUDA-Support

1. Download and unpack OpenMPI 1.7.2 (features)

# wget http://www.open-mpi.org/software/ompi/v1.7/downloads/openmpi-1.7.2.tar.gz
# tar -zxvf openmpi-1.7.2.tar.gz
# cd openmpi-1.7.2

2. Configure the OpenMPI with CUDA Support

# ./configure --prefix=/usr/local/openmpi-1.7.2-intel-cuda CC=icc CXX=icpc F77=ifort FC=ifort --with-cuda=/opt/cuda --with-cuda-libdir=/usr/lib64
# make -j 8
# make install

References:

  1. 34. How do I build Open MPI with support for sending CUDA device memory?