- Ubuntu 16.04
- CUDA 8.0 installed from TensorFlow instructions.
- numba==0.29.0 installed via pip.
NUMBAPRO_CUDALIB
is undocumented and buried in numba/cuda/cudadrv/libs.py
to find.
My numba cuda checking scripts:
check.sh
import serial | |
import sys | |
import datetime | |
port = sys.argv[1] | |
ser = serial.Serial(port=port, baudrate=115200, timeout=5) | |
try: | |
while 1: | |
data = "{},{},{}".format(datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S.%f"), | |
port, |
NUMBAPRO_CUDALIB
is undocumented and buried in numba/cuda/cudadrv/libs.py
to find.
My numba cuda checking scripts:
check.sh
# CUDA. | |
apt-get install wget | |
wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_8.0.44-1_amd64.deb | |
dpkg -i cuda-repo-ubuntu1604_8.0.44-1_amd64.deb | |
apt-get update -y | |
# Recomends on a bare system installs a *ton* of 'junk', including the Ubuntu desktop. debtree chokes on telling you why. | |
apt-get install --no-install-recommends cuda | |
# NVIDIA CUDA® Deep Neural Network library (cuDNN) | |
# https://developer.nvidia.com/cud |
#!/usr/bin/env bash | |
# Config | |
MAINTAINER="Jed Frey" | |
VERSION=3.2.0 | |
# Adapted from http://www.pyimagesearch.com/2016/10/24/ubuntu-16-04-how-to-install-opencv/ | |
apt-get install --yes libjpeg-dev libtiff5-dev libjasper-dev libpng12-dev libavcodec-dev libavformat-dev \ | |
libswscale-dev libxvidcore-dev libx264-dev libv4l-dev liblapacke-dev libgtk-3-dev \ | |
libopenblas-dev libhdf5-dev libtesseract-dev libleptonica-dev \ |
def main(): | |
print("{:3s} {:7s} {:s} {:s}".format("Dec","Float","hex","Hex")) | |
for count in range(0,256): | |
print("{0:3d} {0:5.1f} 0x{0:02x} 0X{0:02X}".format(count)) | |
# Main body | |
if __name__ == '__main__': | |
main() |
## Python3 Caffe Makefile.config | |
# Supports virtualenv | |
## Refer to http://caffe.berkeleyvision.org/installation.html | |
# Contributions simplifying and improving our build system are welcome! | |
# cuDNN acceleration switch (uncomment to build with cuDNN). | |
USE_CUDNN := 1 | |
# CPU-only switch (uncomment to build without GPU support). |
Installing Numba from pip fails since Ubuntu 16.04 only has up through llvm-3.8 in its repositories.
As of version 0.16.0, llvmlite requires LLVM 3.9. It does not support earlier or later versions of LLVM.
Install LLVM-3.9 through the LLVM's Debian/Ubuntu packages (http://apt.llvm.org/).
Add:
@echo off | |
REM This batch file is a short example of how to call | |
REM MATLAB from the command line. | |
REM Options: | |
REM -nosplash Don't display the MATLAB Splash screen | |
REM -wait Don't return to the script until MATLAB is closed. | |
REM -c explicitly specify the license file. | |
REM -logfile Log all script output to the logfile. | |
REM -r Run the specified matlab commands. |
service = com.mathworks.toolbox_packaging.services.ToolboxPackagingService; | |
configKey = service.openProject(project_name); |
VERS="2.4.6 2.5.6 2.6.9 2.7.14 3.0.1 3.1.5 3.2.6 3.3.7 3.4.7 3.5.4 3.6.3 3.7.0a1" | |
for VER in VERS | |
do | |
URL = https://www.python.org/ftp/python/${VER}/Python-${VER}.tgz | |
echo $URL | |
curl -L ${URL} | tar xzv | |
cd Python-${VER} | |
./configure --prefix=/opt/python |