Xiaotong Zhuang
Abstract
Because of stringent power constraints, aggressive latency-hiding approaches, such as prefetching, are absent in the state-of-the-art embedded processors. There are two main reasons that make prefetching power inefficient. First, compiler-inserted prefetch instructions increase code size and, therefore, could increase I-cache power. Second, inaccurate prefetching (especially for hardware prefetching) leads to high D-cache power consumption because of useless accesses. In this work, we show that it is possible to support power-efficient prefetching through bit-differential offset assignment. We target the prefetching of relocatable stack variables with a high degree of precision. By assigning the offsets of stack variables in such a way that most consecutive addresses differ by 1 bit, we can prefetch them with compact prefetch instructions to save I-cache power. The compiler first generates an access graph of consecutive memory references and then attempts a layout of the memory locations in the smallest hypercube. Each dimension of the hypercube represents a 1-bit differential addressing. The embedding is carried out in as compact a hypercube as possible in order to save memory space. Each load/store instruction carries a hint regarding prefetching the next memory reference by encoding its differential address with respect to the current one. To reduce D-cache power cost, we further attempt to assign offsets so that most of the consecutive accesses map to the same cache line. Our prefetching is done using a one entry line buffer [Wilson et al. 1996]. Consequently, many look-ups in D-cache reduce to incremental ones. This results in D-cache activity reduction and power savings. Our prefetcher requires both compiler and hardware support. In this paper, we provide implementation on the processor model close to ARM with small modification to the ISA. We tackle issues such as out-of-order commit, predication, and speculation through simple modifications to the processor pipeline on noncritical paths. Our goal in this work is to boost performance while maintaining/lowering power consumption. Our results show 12% speedup and slight power reduction. The runtime virtual space loss for stack and static data is about 11.8%.
- ARM Co.Ltd, ARM 7TDMI Data Sheet.Google Scholar
- ARM Co.Ltd, ARM7500FE Data Sheet.Google Scholar
- Aho, A. V., Sethi, R., and Ullman, J. D. 1986. Compilers Principles, Techniques and Tools, Addison-Wesley, Reading, MA. Google Scholar
- Brooks, D., Tiwari, V., and Martonosi, M. 2000. Wattch: A framework for architectural-level power analysis and optimizations. ISCA'00 (June). Google ScholarCross Ref
- Basu, K., Choudhary, A., Pisharath, J., and Kandemir, M. 2002. Power protocol: Reducing power dissipation on off-chip data buses. MICRO (Nov.). Google Scholar
- Burger, D. and Austin, T. M. 1997. The SimpleScalar tool set version 2.0. Tech. Report 1342, Univ. of Wisconsin--Madison (May).Google Scholar
- Calder, B., Krintz, C., John, S., and Austin, T. 1998. Cache-conscious data placement. In Proceedings of Architectural Support for Programming Languages and Operating Systems. (Oct). Google Scholar
- Luk Chi-Keung and Mowry, T. C. 1996. Compiler based prefetching for recursive data structures. In Proceedings of the Seventh International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS VII), 222--233 (Oct.). Google Scholar
- Cho, S., Yew, P. C., and Lee, G. 1999. Decoupling local variable accesses in a wide-issue superscalar processor. ISCA, (May). Google Scholar
- Guthaus, M. R., Ringenberg, J. S., Ernst, D., Austin, T. M., Mudge, T., and Brown, R. B. 2001. MiBench: A free, commercially representative embedded benchmark suite. IEEE 4th Annual Workshop on Workload Characterization. Google Scholar
- Haber, G., Klausner, M., Eisenberg, V., Mendelson, B., and Gurevich, M. 2003. Optimization opportunities created by global data reordering. In Proceedings of International Symposium on Code Generation and Optimization. Google Scholar
- Intel Corp. SA-110 Microprocessor Tech. Ref. Manual.Google Scholar
- Lee, H. S., Smelyanskiy, M., Newburn, C. J., and Tyson, G. S. 2001. Stack value file: Custom microarchitecture for the stack. HPCA-7, (Jan). Google Scholar
- Leupers, R. and Marwedel, P. 1996. Algorithm for address assign-ment in DSP code generation. In Proc. ICCAD. Google Scholar
- Liao, S., Devadas, S., Keutzer, K., Tjiang, S., and Wang, A. 1996. Storage assignment to decrease code size. ACM TOPLAS 18, 3, (May), 235--253. Google Scholar
- Noth, W. and Kolla, R. 1990. Spanning tree-based state encoding for low-power dissipation. DATE, 168--174.Google Scholar
- Lipasti, M. H., Schmidt, W. J., Kunkel, S. R., and Roediger, R. R. 1995. Spaid: Software prefetching in pointer-and call-intensive environments. In Proceedings of the 28th Annual IEEE/ACM International Symposium on Microarchi-tecture (MICRO 28), (Nov.). 231--236. Google Scholar
- Ozawa, T., Kimura, Y., and Nishiza-ki, S. 1995. Cache miss heuristics and preloading techniques for general-purpose programs. In Proceedings of the 28th Annual IEEE/ACM International Symposium on Microarchitecture. (MICRO 28) (Nov.). Google Scholar
- Panda, P. R. and Dutt, N. D. 1999. Low-power memory mapping through reducing address bus activity. IEEE Transactions on VLSI Systems 7, 3 (Sept.). Google ScholarDigital Library
- Perez, D. G., Mouchard, G., and Temam, O. 2004. MicroLib: A case for the quantitative comparison of micro-architecture mechanisms. MICRO 37. Google Scholar
- Pomerene, J., Puzak, T., Rechtschaffen, R., and Sparacio, F. 1989. Prefetching system for a cache having a second directory for sequentially accessed blocks. U. S. Patent number 4,807,110 (Feb.).Google Scholar
- Rao, A. and Pande, S. 1999. Storage assignment optimizations to generate compact and efficient code on embedded DSPs. In ACM (PLDI). 128--138. Google Scholar
- Segars, S. 2001. Low power design techniques for microprocessors. ISSCC (Feb.).Google Scholar
- Smith, A. J. 1982. Cache memories. Computing Surveys. 14, 3 (Sept.). Google ScholarDigital Library
- Udayanarayanan, S. and Chakrabarti, C. 2001. Address code generation for DSPs. DAC.Google Scholar
- Wagner, A. and Corneil, D. G. 1990. Embedding trees in a hypercube is NP-complete. SIAM J. Computing 18, 3, 570--590. Google ScholarDigital Library
- Witchel, E., Larsen, S., Ananian, C. S., and Asanovi'c, K. 2001. Direct addressed caches for reduced power consumption. MICRO'01 (Dec.). Google Scholar
- Wilson, K. M., Olukotun, K., and Rosenblum, M. 1996. Increasing cache port efficiency for dynamic superscalar microprocessors. ISCA. Google Scholar
- Wilton, S. and Jouppi, N. P. 1993. An enhanced access and cycle time model for on-chip caches. Technical Report TN93/5, Compaq Western Research Lab.Google Scholar
- Zhuang, X., Lau, C., and Pande, S. 2003. Storage assignment optimizations through variable coalescence for embedded processors. In Proceedings of ACM SIGPLAN Conference on Languages, Compiler, and Tools for Embedded Systems (LCTES-03). Google Scholar
- Zhuang, X. and Pande, S. 2004. Power-efficient prefetching via bit-differential offset assignment on embedded processors. In Proceedings of ACM SIGPLAN Conference on Languages, Compiler, and Tools for Embedded Systems (LCTES-04) (June). Google Scholar
Index Terms
- Power-efficient prefetching for embedded processors
Recommendations
Power-efficient prefetching via bit-differential offset assignment on embedded processors
LCTES '04: Proceedings of the 2004 ACM SIGPLAN/SIGBED conference on Languages, compilers, and tools for embedded systemsDue to stringent power constraints, aggressive latency hiding approaches such as prefetching are absent in the state-of-the-art embedded processors. There are two main reasons that cause prefetching to be power inefficient. First, compiler inserted ...
Power-efficient prefetching via bit-differential offset assignment on embedded processors
LCTES '04Due to stringent power constraints, aggressive latency hiding approaches such as prefetching are absent in the state-of-the-art embedded processors. There are two main reasons that cause prefetching to be power inefficient. First, compiler inserted ...
Efficient Integration of Compiler-Directed Cache Coherence and Data Prefetching
Cache coherence enforcement and memory latency reduction and hiding are very important and challenging problems in the design of large-scale distributed shared-memory (DSM) multiprocessors. We propose an integrated approach to solve these problems ...
Comments