Best of EuroPar '14
Post on 1-9-2014 by CNugteren Tags: GPU, programming, compiler, conference

This year’s EuroPar was held in Porto, a city on the mouth of the river Douro in the north of Portugal.

The conference started with 2 full days of workshops, including the 12th HeteroPar, the 7th MuCoCoS, and the 7th UCHPC. Some related highlights of the workshops are:

  • “A visual programming model to implement coarse-grained DSP applications on parallel and heterogeneous clusters”. An image-processing language to create flows of kernels.
  • “An Empirical Evaluation of GPGPU Performance Model”. A summary of several existing GPU models, including a couple of test cases.
  • A Study of the Potential of Locality-Aware Thread Scheduling for GPUs. This is my own work on optimising thread scheduling for multi-threaded architectures such as the GPU.
  • “Exploiting Hidden Non-uniformity of Uniform Memory Access on Manycore CPUs”. The point of this work was to demonstrate the non-uniformity effects in the Xeon Phi co-processors. Some effects where not fully understood.

The main program included the following interesting talks:

Oh, and here is a picture of my presentation: