论文信息 - Study of combining GPU/FPGA accelerators for high-performance computing

Study of combining GPU/FPGA accelerators for high-performance computing

This contribution presents the performance modeling of a super desktop with GPU and FPGA accelerators, using OpenCL for the GPU and a high-level synthesis compiler for the FPGAs. The performance model is used to evaluate the different high-level synthesis optimizations, taking into account the resource usage, and to compare the compute power of the FPGA with the GPU

[1] Ulrich Brunsmann,et al. FPGA-GPU architecture for kernel SVM pedestrian detection , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Workshops.

[2] E. E. Grant,et al. Exploratory experimental studies comparing online and offline programming performance , 1968, CACM.

[3] Samuel Williams,et al. Roofline: an insightful visual performance model for multicore architectures , 2009, CACM.

[4] Ra Inta,et al. The "Chimera": An Off-The-Shelf CPU/GPGPU/FPGA Hybrid Computing Platform , 2012, Int. J. Reconfigurable Comput..

[5] James Demmel,et al. A view of the parallel computing landscape , 2009, CACM.

[6] Ian Reid,et al. fastHOG – a real-time GPU implementation of HOG , 2011 .