isabella232 / conda-recipe-cudatoolkit-dev

conda recipe for a "full" cudatoolkit package

conda-recipe-cudatoolkit-dev

conda recipe for a "full" cudatoolkit package

Build and Install the package

conda create -n cudatest python=3.6 conda-build numba conda-verify -y

source activate cudatest 

conda build condarecipe9.2/

conda install --use-local $CONDA_PREFIX/conda-bld/linux-64/cudatoolkit-dev-9.2-0.tar.bz2 -v

Testing the package

To confirm that the compiler was properly installed, run:

which nvcc

This returns the installation path

/home/ubuntu/miniconda3/envs/cudatest/bin/nvcc  

To find the compiler's version, run:

nvcc --version

You get:

vcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Tue_Jun_12_23:07:04_CDT_2018
Cuda compilation tools, release 9.2, V9.2.148

This repo contains both python (numba) tests and cuda tests.

To run python (numba) tests, run:

cd tests
python run_test.py

If everything is configured properly, you get:

Finding cublas
	located at /home/ubuntu/miniconda3/envs/cudatest/lib/libcublas.so.9.2.148
	trying to open library...	ok
Finding cusparse
	located at /home/ubuntu/miniconda3/envs/cudatest/lib/libcusparse.so.9.2.148
	trying to open library...	ok
Finding cufft
	located at /home/ubuntu/miniconda3/envs/cudatest/lib/libcufft.so.9.2.148
	trying to open library...	ok
Finding curand
	located at /home/ubuntu/miniconda3/envs/cudatest/lib/libcurand.so.9.2.148
	trying to open library...	ok
Finding nvvm
	located at /home/ubuntu/miniconda3/envs/cudatest/lib/libnvvm.so.3.2.0
	trying to open library...	ok
	finding libdevice for compute_20...	ok
	finding libdevice for compute_30...	ok
	finding libdevice for compute_35...	ok
	finding libdevice for compute_50...	ok
NVVM version (1, 5)

To run CUDA (Thrust) tests, run:

nvcc -o version thrust_version.cu
./version

You get:

Thrust v1.9