High-Performance Physical-Independent Address-Based Communication Interface for FPGA in Custom Scientific Equipment

Corna, N.; Ronconi, E.; Garzetti, F.; Salgaro, S.; Lusardi, N.; Tavazzani, L.; Geraci, A.

doi:10.1109/nss/mic42677.2020.9507844

Cited by 2 publications

(2 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When compared to similar COBS, Multi-COBS can be utilized to frame packets exchanged with such systems, providing the aforementioned benefits: very low average and worst-case overhead and high throughput. In fact, [11] designed and documented a communication system that leverages Multi-COBS for framing. A communication system like this was used to connect an x86_64 workstation to a high-performance FPGA-based Time-todigital converter (TDC) measuring device [9], [25] through multiple physical links such UART, USB 2, and USB 3.…”

Section: Real Applicationsmentioning

confidence: 99%

“…Common scenarios include PL devices that send data to a workstation where they are further elaborated [8], [9] or PL devices that can be configured from software running on a computer [7]. For such reason, a wide number of communication frameworks for PLmicroprocessor links have been recently developed [10], [11]. Such systems rely on framing techniques that should be capable of being executed on both PL and microprocessorbased devices.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-COBS: A Novel Algorithm for Byte Stuffing at High Throughput

et al. 2022

View full text Add to dashboard Cite

Framing methods are used to break a data stream into packets in most digital communications. The use of a reserved symbol to denote the frame boundaries is a popular practice. This end-of-frame (EOF) marker should be removed from the packet content in a reversible manner. Many strategies, such as the bit and byte stuffing processes employed by high-level data link control (HDLC) and Point-to-Point Protocol (PPP), or the Consistent Overhead Byte Stuffing (COBS), have been devised to perform this goal. These bit and byte stuffing algorithms remove the reserved EOF marker from the packet payload and replace it with some extra information that can be used to undo the action later. The amount of data added is called overhead and is a figure-of-merit of such algorithms, together with the encoding and decoding speed. Multi-COBS, a new byte stuffing algorithm, is presented in this paper. Multi-COBS provides concurrent encoding and decoding, resulting in a performance improvement of factor four or eight in common word-based digital architectures while delivering an average and worst-case overhead equivalent to the state-of-the-art. On the reference 28-nanometer field programmable gate array (FPGA) (Artix-7), Multi-COBS achieves a throughput of 6.6 Gbps, instead of 1.7 Gbps of COBS. Thanks to its parallel elaboration capability, Multi-COBS is ideal for digital systems built in programmable logic as well as modern computers.

show abstract

Section: Real Applicationsmentioning

confidence: 99%