[eigen] Help with Eigen on SYCL 1.2

Hi all,

We at Codeplay are looking for a student who would be interested in a paid internship that aims to develop a SYCL back-end for Eigen.

More specific description below:

Explanation of what the intern would be working on

The main focus of the project is to be able to provide an interface that can work with SYCL alongside CUDA, Thrust or any other parallel programming models, recognising the capabilities and restrictions of GPGPU models in correlation to highly optimised CPU algorithms and interfaces. Subsequently, examples will be produced using that interface that will work with existing or trivially altered examples of the Eigen library, showing the advantages and drawbacks of using such an interface. Performance comparisons between the CPU and GPUGPU samples of the Eigen library will be a driving factor for the viability and continuation of this work by the Eigen community.

As a contributor to Eigen, the student will be working closely with the active developers in its community, as well as the contributors to the SYCL specification and the first announced implementation of it, Codeplay's ComputeCpp system. Support and help from the ComputeCpp team will be available to the student, as well as from developers who have been working on research projects, demos and benchmarks for SYCL and OpenCL.

Essential Skills required of the intern (if any)

Good knowledge of C++ is required, in particular templates, STL and preferably C++11

Experience with Eigen

Some GPGPU programming (CUDA or OpenCL)

Length of proposed internship

6 months, possibly extended to 12 months.

For any questions don't hesitate to send me an email.

Thanks,