Designing a high performance and low energy-consuming embedded system with considering code compressed environments

In this paper, we present an embedded system design which is cost-efficient and which considers performance improvement and power consumption based on the frequency with which instructions are executed. We use the locality of running programs to optimize the use of memory space, system performance, and power consumption; that is, we compress infrequently executed codes to save the use of memory space but compress (encode) frequently executed codes to power consumption and maximize performance. To save the number of memory access times, the frequently executed instructions are encoded as shorter code words and then the continuous code words are packed into a pseudo instruction. Once the decompression engine fetches one pseudo instruction, it can extract multiple instructions to reduce the number of memory access times. In addition, we also propose a design with multiple look-ahead tables that solves the problem which arises due to the reduction in instructions from one-time memory access which occurs because the great amount of frequently executed instructions. From our simulation results, our method with one 256-instruction look-ahead table does not increase the compression ratio, and the ratio of the power consumption can be reduced by about 47.23% than compressing all instructions. With regard to multiple look-ahead table methods, when one 512-instruction look-ahead table is used, the ratio of the power consumption is reduced by 37.25% only, however, when two look-ahead tables that contain 256 instructions in each table are used, the power consumption can be improved by 49.36%. According to the simulation results, our proposed methods based on the frequencies of executed instructions result in low power consumption, performance improvement and reduced memory space.