hughperkins / tf-coriander

OpenCL 1.2 implementation for Tensorflow

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Tensorflow-cl

Run Tensorflow on OpenCL™ 1.2 devices

  • tested on:
    • Mac Sierra using Radeon Pro 450 GPU (thank you ASAPP :-) )
    • Ubuntu 16.04, using NVIDIA K520
  • should work theoretically on any OpenCL 1.2 GPU

Piccie

On a Mac:

Execution speed

What's working

Installation

Tests

Design/architecture

  • tensorflow code stays 100% NVIDIA® CUDA™
  • Coriander compiles the NVIDIA® CUDA™ code into OpenCL
  • Cedric Nugteren's CLBlast provides BLAS (matrix multiplications)

Presentation on Coriander at this year's IWOCL 2017

Related projects

News

About

OpenCL 1.2 implementation for Tensorflow

License:Apache License 2.0


Languages

Language:C++ 46.0%Language:Python 41.6%Language:Jupyter Notebook 6.2%Language:TypeScript 2.3%Language:HTML 1.7%Language:Shell 0.9%Language:C 0.3%Language:CMake 0.2%Language:Objective-C++ 0.2%Language:Java 0.2%Language:Go 0.2%Language:Makefile 0.1%Language:JavaScript 0.0%Language:Objective-C 0.0%Language:CSS 0.0%