Reconfigurable Parallel Computing Architecture for On-Board Data Processing

Abstract

The increasing use of new Earth-observation and communication technologies coupled with rapidly changing customer needs require a high performance and flexible processing technology for on-board data processing. Currently available processors for space cannot handle processing such a large magnitude. An alternative approach would be the development of application specific ICs (ASICs) which offer high processing power, but the lack of flexibility for adaptation to emerging standards and improvements is a major disadvantage. The extreme processing platform (XPP) developed by PACT AG offers a solution to this dilemma by providing the flexibility of a processor kernel along with the performance similar to an ASIC. The XPP is a runtime reconfigurable data processing technology built using a scalable array of arithmetic processing units, embedded memories, high bandwidth I/O, a packet oriented internal network and designed to support parallelism. This paper gives an overview of the XPP core architecture, compares the technology with currently available space processors, and assesses the implementation of wavelet transformation algorithm on XPP and the transfer of XPP architecture to a radiation tolerant semiconductor technology