Objective
This talk not only focus on pyCUDA but also on codepy and Boost Python which made CUDA possible to integrate in python.
Description
This talk starts with introduction to modules in python and how to implement them. Then, Boost python which helps codepy to implement metaprogramming in python. Then, on pyCUDA, how all the delivered talk gets together to form such a powerful package and implementation of pyCUDA. Implementation of Copperhead, the total scripting module for GPU.
Requirements
Linux running NVIDIA GPU with new drivers and CUDA 5.5 install
Speaker bio
I am a computer geek. I love python.
My focus are parallel programming, compilers for multi-cored processors.
I program for fun. Develop search engines, websites.
Currently working for Dirt9.
2
▼
Parallel programming using python. New concept and technology. Looking forward to it.
1
▼
It is interesting to see talks on GPUs and CUDA and I would like to + vote this. But there is no meat in the talk description and nothing about the author. Please include more data here.
1
▼
Hi, I was unable to edit the description. Don't mind, I am expalining it here.
The talk mainly focus on building of modules. The development of modules is made easy by Boost:Python. My talk focus on how to use Boost Python. Then, I'll talk about codepy. This is a metaprogramming package for python which builds and import modules automatically to python. Then, on CUDA and OpenCL. How to use these in Python to achieve HPC. pyCUDA and pyCL is the scripting package for python. If time permits I love to talk about Copperhead, a scripting HPC module for python.
About me: I got selected to give a talk in Pycon Canada, Tokyo, Poland this year. I am a python, c geek. My focus is HPC.
Links:
https://developer.nvidia.com/content/copperhead-data-parallel-python
http://mathema.tician.de/software/pycuda
http://mathema.tician.de/software/pyopencl
http://mathema.tician.de/software/codepy
My git repo:
https://github.com/adityaavinash