Adaptive Execution of Compiled Queries

Compiling queries to machine code is arguably the most efficient way for executing queries. One often overlooked problem with compilation, however, is the time it takes to generate machine code. Even with fast compilation frameworks like LLVM, Generating machine code for complex queries routinely takes hundreds of milliseconds. Such compilation times can be a major disadvantage for workloads that execute many complex, but quick queries. To solve this problem, we propose an adaptive execution framework, which dynamically and transparently switches from interpretation to compilation. We also propose a fast bytecode interpreter for LLVM, which can execute queries without costly translation to machine code and thereby dramatically reduces query latency. Adaptive execution is dynamic, fine-grained, and can execute different code paths of the same query using different execution modes. Our extensive evaluation shows that this approach achieves optimal performance in a wide variety from settings---low latency for small data sets and maximum throughput for large data sizes