Parallel Distributed Infrastructure for Minimization of Energy

ParaDIME at the 23rd IEEE International Symposium on Field-Programmable Custom Computing Machines

Thu, 2015-05-07
Researcher Oriol Arcas presented the ParaDIME poster "Heterogeneous Platform to Accelerate Compute Intensive Applications" at the 23rd IEEE International Symposium on Field-Programmable Custom Computing Machines which took place from the 3rd to the 5th of May in Vancouver (Canada).
 
This symposium is the original and premier forum for presenting and discussing new research related to computing that exploits the unique features and capabilities of FPGAs and other reconfigurable hardware. Over the past two decades, FCCM has been the place to present papers on architectures, tools, and programming models for field-programmable custom computing machines as well as applications that use such systems.
 
The work presented at Vancouver was a novel heterogeneous acceleration platform to accelerate computing-intensive applications like face recognition. Such platform not only could greatly improve the performance, but it also could reduce the energy consumption. This was possible thanks to the interconnection of two advanced system-on-chip accelerators: an NViDIA Tegra platform, which includes a CPU and a GPU, and a Xilinx ZYNQ platform, which combines a CPU and an FPGA. All the four components accelerated different parts of the algorithm, achieving better results than each component alone.