Code compression for VLIW processors using variable-to-fixed coding

Abstract

Memory has been one of the most restricted resources in the embedded computing system domain. Code compression has been proposed as a solution to this problem. Previous work used fixed-to-variable coding algorithms that translate fixed-length bit sequences into variable-length bit sequences. In this paper, we propose code compression schemes that use variable-to-fixed (V2F) length coding. We also propose an instruction bus encoding scheme, which can effectively reduce the bus power consumption. Though the code compression algorithm can be applied to any embedded processor, it favors VLIW architectures because VLIW architectures require a high-bandwidth instruction pre-fetch mechanism to supply multiple operations per cycle. Experiments show that the compression ratios using memoryless V2F coding for IA-64 and TMS320C6x are around 72.7% and 82.5% respectively. Markov V2F coding can achieve better compression ratio up to 56% and 70% for IA-64 and TMS320C6x respectively. A greedy algorithm for codeword assignment can reduce the bus power consumption and the reduction depends on the probability model used.