Title :
A Hybrid FPGA Model to Estimate On-Chip Crossbar Logic Utilization in SoC Platforms
Author :
Yoon Kah Leow ; Akoglu, Ali
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Arizona, Tucson, AZ, USA
Abstract :
FPGA analytical models, that express the relationship between architectural parameters (e.g., LUT size, cluster size, inputs per cluster, etc) and performance (e.g., logic utilization, critical path delay, power, etc), have been designed mainly for island-style FPGAs targeting a general application. Therefore, most analytical models will produce inaccurate results when heterogeneous FPGA architectures are targeted. Furthermore, the inherent continuous nature of mathematical models also prevent them from capturing the discrete effects of uniform circuits. Example of such circuits are crossbar switches and barrel shifters. In this paper, we derive a biased model that captures the discrete effects with respect to the logic utilization of crossbar switches by varying the LUT size.
Keywords :
computer architecture; field programmable gate arrays; switches; system-on-chip; FPGA analytical model; LUT size; SoC platforms; architectural parameters; barrel shifters; crossbar switch logic utilization; discrete effects; heterogeneous FPGA architecture; hybrid FPGA model; island-style FPGA; mathematical models; on-chip crossbar logic utilization estimation; system-on-chip platforms; uniform circuits; Analytical models; Field programmable gate arrays; Integrated circuit modeling; Mathematical model; Multiplexing; System-on-chip; Table lookup; crossbar switches; discrete effects; logic utilization;
Conference_Titel :
Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW), 2013 IEEE 27th International
Conference_Location :
Cambridge, MA
Print_ISBN :
978-0-7695-4979-8
DOI :
10.1109/IPDPSW.2013.138