PLC 2015 - Workshop 2015 Programming Models, Languages and Compilers for Manycore and Heterogeneous Architectures

Three essential pillars to successful parallel computing - Productivity, Portability, Performance. Creating software for heterogeneous systems can be quite complex especially when the low-level details need to be managed and abstracted from the programmer. Emerging standards are providing an incremental development to target heterogeneous architectures, be it NVIDIA, ARM, Intel or AMD. We all know software is an expensive investment. Portability is necessary, ensuring long lifetime of the software and thus reducing the maintenance cost. Other challenges include locality and memory issues, load balancing, hiding latency with concurrency and so on.
This workshop aims to brainstorm ways to make programming heterogeneous systems less challenging and more interesting. We believe that this workshop will provide a forum for the presentation and discussions of research on all aspects of heterogeneous systems programming models, compiler optimizations, language extensions, and software tools for such systems.
Areas of interest include but are not limited to the following topics:
Multicore processors, Heterogeneous systems, Accelerators
Programming models: thread and task based models, data parallel models, stream programming
Programming environments for heterogeneous systems
OpenMP for Accelerator
OpenACC
OpenCL/CUDA
OpenHMPP and other related models
DSL
Compiler optimizations and tuning heterogeneous systems
Parallelization and loop transformations
Locality optimizations
SIMDization/Vectorization
Reducing synchronization and scheduling overheads on heterogeneous systems
Tiling, parametric tiling and offloading
Runtime systems for multicore processors, heterogeneous systems, accelerators
Debuggers, and performance analysis tools for heterogeneous system and accelerators
Application and Benchmarks