Skip to content

Illustrates how to build Coriander (CUDA to OpenCL translation tool) with Oclgrind (OpenCL debugger) and AIWC (for Workload Characterization)

Notifications You must be signed in to change notification settings

ANU-HPC/coriander-and-oclgrind

Repository files navigation

This repository was built to illustrate how Coriander and Oclgrind can be used together. Unfortunately, the devil is in the details, -- the set-up is painful -- since it requires both libraries to be dynamically linked against the same version of LLVM to operate to avoid the following error:

error "'phi-node-folding-threshold' registered more than once"

This repository uses Docker to outline the build steps I used to get these two awesome tools to play nice together.

Dependencies

Precanned

You can avoid building the project by using my prebuilt image. This is done with the following command:

docker run --rm -it --runtime=nvidia beaujoh/coriander-and-oclgrind:1.0

If you are feeling like being more explicit, you can build and run directly outlined in the remainder of the README.

Build

docker build -t coriander-and-oclgrind .

Run

docker run --rm -it --runtime=nvidia coriander-and-oclgrind /bin/bash
make test

About

Illustrates how to build Coriander (CUDA to OpenCL translation tool) with Oclgrind (OpenCL debugger) and AIWC (for Workload Characterization)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published