首页> 外文会议>International Conference on Emerging Trends in Communication, Control, Signal Processing and Computing Applications >Miss rate analysis of cache oblivious matrix multiplication using sequential access recursive algorithm and normal multiplication algorithm
【24h】

Miss rate analysis of cache oblivious matrix multiplication using sequential access recursive algorithm and normal multiplication algorithm

机译:高速缓存透露矩阵乘法使用顺序访问递归算法和正常乘法算法的错过率分析

获取原文

摘要

Cache oblivious algorithms are designed to get the good benefit from any of the underlying hierarchy of caches without the need to know about the exact structure of the cache. These algorithms are cache oblivious i.e., no variables are dependent on hardware parameters such as cache size and cache line length. Optimal utilization of cache memory has to be done in order to get the full performance potential of the hardware. We present here the miss rate comparison of cache oblivious matrix multiplication using the sequential access recursive technique and normal multiplication program. Varying the cache size the respective miss rates in the L1 cache are taken and then comparison is done. It is found that the miss rates in the L1 cache for the cache oblivious matrix multiplication program using the sequential access recursive technique is comparatively lesser than the naive matrix multiplication program.
机译:缓存忽略算法旨在从任何底层的高速缓存的底层层次结构获得良好的益处,而无需了解缓存的确切结构。这些算法是缓存忽略的i.e.,没有变量取决于硬件参数,例如高速缓存大小和高速缓存行长度。必须完成高速缓冲存储器的最佳利用,以便获得硬件的完整性能潜力。我们在这里介绍了使用顺序访问递归技术和正常乘法程序的缓存透露矩阵乘法的错过率比较。改变缓存大小,拍摄L1缓存中的相应未命中率,然后进行比较。结果发现,使用顺序访问递归技术的高速缓存透露矩阵乘法程序的L1缓存中的错过率比天真矩阵乘法程序相对较低。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号