Methodology for Refinement and Optimisation of Dynamic Memory Management for Embedded Systems in Multimedia Applications

Abstract

In multimedia applications, run-time memory management support has to allow real-time memory de/allocation, retrieving and processing of data. Thus, its implementation must be designed to combine high speed, low power, large data storage capacity and a high memory bandwidth. In this paper, we assess the performance of our new system-level exploration methodology to optimise the memory management of typical multimedia applications in an extensively used 3D reconstruction image system [1, 2]. This methodology is based on an analysis of the number of memory accesses, normalised memory footprint 1 and energy estimations for the system studied. This results in an improvement of normalised memory footprint up to 44.2% and the estimated energy dissipation up to 22.6% over conventional static memory implementations in an optimised version of the driver application. Finally, our final version is able to scale perfectly the memory consumed in the system for a wide range of input parameters whereas the statically optimised version is unable to do this.

Application-Layer Fault-Tolerance Protocols

“…The initial version in C, based on the C++ development code was not even close to be able to process the images in soft real time and only 7 fps were processed. In a following step, the algorithm was optimized by including faster and more memory aware data types and by reducing intermediate buffers and loop transformations (Leeman et al, 2005). Due to these changes, this algorithm got a speedup of 30%, which still put it about two times slower than the minimal soft real time requirement (10-12 fps).…”

“…The default filter leaves the expressions unmodified, but the programmer can choose otherwise. This simple idea translates in a powerful tool for the programmer: For instance (Shigeru, 1996), shows how easy it is to perform some accounting on all methods invocations of a given class (doing this in a standard programming language such as C++ would require considerable programming effort (Leeman et al, 2005)). …”

Systematic intermediate sequence removal for reduced memory accesses

“…Assuming that the intermediate buffer is then only used once, which is true for multimedia applications working in phases (as observed in [12]), all the writes to inter as well as the subsequent reads from inter will be completely removed. In this simple motivational example, removing inter should therefore lead to a reduction in memory accesses of up to 66%.…”

Dynamic memory management design methodology for reduced memory footprint in multimedia and wireless network applications

“…Even if average values of possible memory footprint estimations are used, these static solutions will result in higher memory footprint figures (i.e. 22% more) than DM solutions [8]. Moreover, these intermediate static solutions will not work in extreme cases of input data, whereas DM solutions can do it.…”

“…For example, in new 3D vision algorithms [15], a well designed custom DM manager can improve memory footprint up to 44.2% over conventional general-purpose DM managers [8].…”