Database Architecture Optimized for the New Bottleneck: Memory Access
VLDB '99 Proceedings of the 25th International Conference on Very Large Data Bases
Weaving Relations for Cache Performance
Proceedings of the 27th International Conference on Very Large Data Bases
Buffering databse operations for enhanced instruction cache performance
SIGMOD '04 Proceedings of the 2004 ACM SIGMOD international conference on Management of data
Improving instruction cache performance in OLTP
ACM Transactions on Database Systems (TODS)
Optimization of frequent itemset mining on multiple-core processor
VLDB '07 Proceedings of the 33rd international conference on Very large data bases
Cache-conscious buffering for database operators with state
Proceedings of the Fifth International Workshop on Data Management on New Hardware
| Hi-index | 0.00 |
The difference between CPU access costs and memory access costs incurs performance degradations on RDBMS. One of the reason why instruction cache misses occur is the size of footprint on RDBMS operations does not fit into a L1 instruction cache. To solve this problem Zhou proposed the buffering operator which changes the order of operation executions. Although the buffering operator is effective, it cannot be applied for RDBMS in the real business. It is because Zhou does not show an algorithm for optimizer to select the buffering operator. Thus we realized the CC-Optimizer which includes an algorithm to appropriately select the buffering operator. Our contributions are the design of new algorithm on an optimizer and its implementation to RDBMS. For experimental RDBMS, we used PostgreSQL as Zhou did, and our machine environment was Linux Kernel 2.6.15, CPU Intel Pentium 4(2.40GHz). The result of preliminary experiments showed that CC-Optimizer was effective. The performance improvement measured by using OSDL DBT-3, was 73.7% in the greatest result, and 17.5% in all queries.