AXI-based SpaceFibre IP Core Implementation SpaceFibre, Poster Paper

被引:0
作者
Jungewelter, D. [1 ]
Cozzi, D. [1 ]
Kleibrink, D. [1 ]
Korf, S. [1 ]
Hagemeyer, J. [1 ]
Porrmann, M. [1 ]
Ilstad, J. [2 ]
机构
[1] Univ Bielefeld, CITEC, Cognitron & Sensor Syst Grp, Inspirat 1, D-33619 Bielefeld, Germany
[2] European Space Agcy, TEC EDP, Estec, NL-2220 AG Noordwijk, Netherlands
来源
PROCEEDINGS OF THE 2014 6TH INTERNATIONAL SPACEWIRE CONFERENCE (SPACEWIRE) | 2014年
关键词
SpaceFibre; SpFi; AXI; FPGA; IP core; DMA; SoC; DRPM;
D O I
暂无
中图分类号
V [航空、航天];
学科分类号
08 ; 0825 ;
摘要
The steadily increasing demand of high-throughput interfaces, e. g., in satellite payload processing systems, drives the development of faster data transmission systems. The emerging SpaceFibre standard offers a multi-gigabit serial connection which is specified on the physical and data link layer while reusing the SpaceWire protocol specification on the higher protocol layers, thus enabling compatibility on the software layer. The AXI SpaceFibre IP core presented in this paper combines the SpaceFibre CODEC IP core developed by STAR-Dundee, a TLK2711 WizardLink Transceiver (meeting the SpaceFibre specification with up to 2.7 Gbit/s), and a DMA interface that connects the IP core to any AXI-based reconfigurable system-on-chip using FPGAs. The IP core configuration and initialization registers for the SpaceFibre RX and TX channels are accessible via AXI4-lite slave interfaces while the payload data is handled by a dedicated scatter/gather AXI-DMA core, which is connected to AXI-Stream FIFOs to provide the maximal possible payload transaction performance. The SpaceFibre IP core can be configured to implement 1 to 8 virtual channels. To evaluate the performance of the SpaceFibre IP core, we integrated it into the " Dynamically Reconfigurable Processing Module" (DRPM), a multi FPGA platform for satellite data payload processing. The AXI-based SpaceFibre IP core was synthesized on a Xilinx Spartan-6 LX150, utilizing in total 2668 slices and 36 BRAMs. For data segments larger than 1 kByte, a bandwidth of approximately 1.9 Gbit/s was achieved, corresponding to 95 % of the possible bandwidth. The IP core will be part of the ESA IP core repository.
引用
收藏
页数:6
相关论文
共 13 条
[1]  
[Anonymous], 2008, ECSSEST5012C
[2]  
ARM, 2010, ARM AMBA AXI PROT VE
[3]  
ESTEC, 2008, SPW RTC AT7913E
[4]  
Florit A. Ferrer, SPACEWIRE C 2013 GOT
[5]  
Hagemeyer J., 2012, ADAPTIVE HARDWARE SY
[6]  
Masuzaki T., SPACEWIRE C 2013 GOT
[7]  
Otake Y., SPACEWIRE C 2013 GOT
[8]  
Parkes Steve, 2012, SPACEFIBRE STANDARD
[9]   RAPTOR - A Scalable Platform for Rapid Prototyping and FPGA-based Cluster Computing [J].
Porrmann, Mario ;
Hagemeyer, Jens ;
Romoth, Johannes ;
Strugholtz, Manuel ;
Pohl, Christopher .
PARALLEL COMPUTING: FROM MULTICORES AND GPU'S TO PETASCALE, 2010, 19 :592-599
[10]  
Texas Instruments, 2008, US MAN TI TLK2711 1