Publication
Bespoke OLAP: Synthesizing Workload-Specific One-size-fits-one Database Engines
Johannes Wehrstein; Timo Eckmann; Matthias Jasny; Carsten Binnig
In: Proceedings of the VLDB Endowment (PVLDB), Vol. 19, No. 11, Pages 3759-3771, arXiv, 2026.
Abstract
Modern OLAP engines support arbitrary analytical workloads, but
this flexibility incurs overhead from runtime schema interpreta-
tion, generic data representations, and abstraction layers, even in
compiled-query systems. Workload-specific engines can eliminate
these costs and exploit specialized data structures and algorithms
for higher performance, yet have historically been too expensive to
build manually. Recent advances in LLM-based code synthesis chal-
lenge this tradeoff, but naive prompting does not produce correct
or efficient engines due to deep architectural dependencies and the
need for systematic refinement. We present Bespoke OLAP, a fully
autonomous synthesis pipeline that constructs high-performance
OLAP engines tailored to a target workload through iterative per-
formance evaluation and automated validation. Bespoke OLAP gen-
erates engines from scratch within minutes to hours and achieves
order-of-magnitude speedups over DuckDB and Umbra, demon-
strating that the generality tax extends beyond query compilation
to storage layout and algorithmic design.