Dynamically increasing the scope of code motions during the high-level synthesis of digital circuits

Abstract

The quality of high-level synthesis results for designs with complex and nested conditionals and loops can be improved significantly by employing speculative code motions. Two techniques are presented that add scheduling steps to the branch of a conditional construct with fewer scheduling steps. This 'balances' or equalises the number of scheduling steps in the conditional branches and increases the scope for application of speculative code motions. These branch balancing techniques have been applied 'dynamically' during scheduling. The authors have implemented algorithms for dynamic branch balancing techniques in a C-to-VHDL high-level synthesis framework called Spark. The utility of these techniques is demonstrated by experimental results on four designs derived from two moderately complex applications, namely, MPEG-1 and the GIMP image processing tool. These results show that the two branch balancing techniques can reduce the cycles on the longest path through the design by up to 38% and the number of states in the controller by up to 37%.

Dynamic Memory Access Management for High-Performance DSP Applications Using High-Level Synthesis

“…This model allows data and conditional semantics to be managed in the same way to maximize the exploitation of parallelism. Well-known data-flow algorithms for the optimization and the synthesis steps can be used with potential management of the control structures to implement mutually exclusive scheduling methods [37], [38].…”

A New ILP Based Approach to Schedule and Bind Simultaneously

“…And the actual functional unit to execute the operations in data-path will be computed at allocation stage. At present, the most of scheduling algorithms will schedule operation firstly, when the scheduling is accomplished, then the binding computation will begin to proceed [5,6,7]. With respect to the clock cycle, the scheduling algorithm need to know the propagation delay of each operation to determine whether two operations can be chained, whether an operation is multi-cycle operation and to determine the rhythm of pipeline operations in data-path [8].…”

Seamless Signal Processing Block Implementation Using the Cubed-C Design Environment

“…It mainly targets digital signal processing (DSP) applications but also more general streaming systems can be included along with specific performance constraints. Mutually exclusive scheduling methods [19][20][21] are implemented with the "Extended Data-flow Graph". This is achieved because EDFG allows for data and conditional semantics to be handled in the same way, and thus the exploitation of potential design parallelism can be maximized.…”