Skip to content
Preprint

The Case for TPU-Accelerated Analytical Query Processing

Oct 2026 · 0 citations · 39 references
Computer Science

TL;DR

This paper presents the first comprehensive case for TPU-accelerated analytical query processing by running all the TPC-H queries entirely on the TPUs, and introduces Native TPU Accelerator (NTA) for analytical query processing, which provides hardware-conscious algorithm design and optimizations in JAX/XLA.

Abstract

Large-scale analytical query processing requires computing substrates capable of immense distributed scale-out. Tensor Processing Units (TPUs) perfectly embody this extreme-scale architecture, offering massive computational density and highly scalable optical pod-level interconnects. Despite these advantages, TPUs have so far remained largely unexplored for analytical database workloads. In this paper, we present the first comprehensive case for TPU-accelerated analytical query processing by running all the TPC-H queries entirely on the TPUs. We introduce Native TPU Accelerator (NTA) for analytical query processing, which provides hardware-conscious algorithm design and optimizations in JAX/XLA. By exposing the TPU's structural mechanics, we demonstrate how dynamic, memory-bound database operations can be rewritten into deterministic JAX programs. We engineer a suite of novel execution strategies that leverage both the TPU's TensorCore and SparseCore to trade sparse, unpredictable memory accesses for dense, hardware-aligned ones. At the core of NTA, a compiler uses catalog statistics to assign guarded static capacities and a hardware-calibrated local cost model to select a TPU-native implementation for each relational operator before JIT compilation. We evaluate our system using the TPC-H benchmark. Taking full advantage of TPU slices and JAX distributed computing, we seamlessly scale our execution environment to run TPC-H benchmark up to Scale Factor 1000 (SF1000). Evaluated against a rigorous CPU ground truth (DuckDB), our native TPU accelerator achieves state-of-the-art performance, achieving a 7.0x single-device speedup over DuckDB at SF100 and a 68.6x speedup with 16 TPUs at SF1000 (1TB). The results establish TPUs not merely as dense neural compute hardware, but as a highly competitive, massively scalable substrate for the next generation of analytical query processing.

View source

Similar papers

Oct 2026

Fast Matrix Multiplication for Database Query Processing

This research drives the core insight of the research: a comprehensive exploration of FMM within database query engines, bringing the theoretical power of algebraic techniques to practical data systems.

Xiao Hu · 0 citations
Preprint Sep 2026

Unifying In-Memory Data Analytics through Sparse Compilation

This work introduces a novel intermediate representation (IR), grounded in relational algebra and sparse iteration theory, that provides a unified abstraction for data and computation that enables workload-agnostic, end-to-end optimizations across diverse analytics applications.

Anand Jayarajan, Gennady Pekhimenko · 0 citations
#artificial intelligence Conference Open access Sep 2026

Efficient Iterative Retrieval with Heterogeneous Batching

Orthrus is a serving system that performs heterogeneous batching within a unified inference loop that addresses the primary challenge lies in unifying embedding and generation workloads with conflicting computational patterns while optimizing batch composition for high performance.

Dohyun Park, Hubertus Franke, Daniel G. Waddington et al. · 1 citation
#artificial intelligence Open access Oct 2026

Trillion-scale integrated framework for high-throughput materials databases and seamless sharing

This framework uses stateless parallel computing with message queues and containerization, and a lakehouse architecture with columnar storage and metadata-based cloud sharing to construct a ten-trillion-scale HEA database in one week.

Xiao-Ya Huang, Yu Liu, Shuo Shi et al. · 0 citations
#artificial intelligence Preprint Oct 2026

Characterizing Parallelism Strategies in LLM Inference: Fundamental Compute-Communication Trade-offs

Large Language Model (LLM) inference has become the dominant workload in modern AI systems, requiring serving infrastructures to maximize throughput while meeting strict latency Service-Level Objectives (SLOs). Since state-of-the-art LLMs exceed the compute and memory capacity of a single GPU, inference is commonly dis...

Javad Mirzaei, Jeebak Mitra · 1 citation · ⚡1
Preprint Oct 2026

Distributed Datalog at Scale with Worst-Case Optimal Joins

Scalable logic-based analytics are frequently dominated by cyclic multi-way joins over large relational datasets. State-of-the-art systems evaluate these joins through chains of binary joins, incurring substantial overhead from large intermediate relations that must be materialized and redistributed across the network....

Kun-Ting Qi, Yi-Hao Sun, Kristopher K. Micinski et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.