Cache-Conscious Query Processing

Foundations A cache is a hardware unit that speeds up access to data. Several cache units may be present at various levels of the memory hierarchy, depending on the processor architecture. For example, a processor may have a small but fast Level-1 (L1) cache for data, and another L1 cache for instructions. The same processor may have a larger but slower L2 cache storing both data and instructions. Some processors may even have an L3 cache. On multicore processors, the lower level caches may be shared among groups of cores. Some initial analysis would typically be performed to determine the performance characteristics of a workload. For example, Ailamaki et al. [1] used hardware performance counters to demonstrate that several commercial systems were, at that time, suffering many L2 data cache misses and L1 instruction cache misses. Based on such observations, one can determine that the L2 data cache and L1 instruction cache are targets for performance tuning. If the operating system does not provide direct access to system parameters such as the cache size, a database system can run a calibration test to estimate the relevant parameters [13]. To get good cache performance, algorithm designers typically utilize one or more of the following general approaches: