Gravity achieves its substantial performance by compiling queries down to raw assembly through a multi-stage compilation process. To find the true ceiling of their hardware, developers benchmarked heavily optimized, Claude-written Rust programs using SIMD against raw columnar storage. Gravity's compiled machine code now runs within 20% of those handwritten baseline tests. Furthermore, the engine retains complete compatibility with standard Postgres query plans, parsers, and native data types.
Key points
The new engine operates within twenty percent of the speed achieved by heavily optimized…: The new engine operates within twenty percent of the speed achieved by heavily optimized handwritten Rust baselines.
Gravity transforms native Postgres query plans into raw assembly using an intermediate pi…: Gravity transforms native Postgres query plans into raw assembly using an intermediate pipeline graph and a Lisp representation.
Full compatibility with the existing Postgres parser, planner, and SQL dialect is preserv…: Full compatibility with the existing Postgres parser, planner, and SQL dialect is preserved without sacrificing low-level execution speed.
Sources checked by our analysis pipeline: 1
With the release of pgrust v0.4-beta, our new query engine, Gravity, is now as much as 3x faster than ClickHouse on ClickBench, the benchmark that ClickHouse themselves designed to show off how fast ClickHouse is. Even better, Gravity achieves that speed while supporting the full Postgres SQL dialect and type system.
In a couple of previous posts, I wrote about how we used batching, operator fusion, and SIMD to make pgrust hundreds of times faster than Postgres for analytical queries, and how we built a copy-and-patch JIT compiler that compiles stencils in 5 microseconds. All of that work, built up over several releases, is what became Gravity.
Copy-and-patch was a great starting point, but gluing pre-compiled templates together only gets us so far. Because we did not have a real intermediate representation or register allocation across the whole loop, we ended up leaving a lot of performance on the table compared to an optimizing compiler like rustc or Clang. For pgrust v0.4, we wanted to go all the way and build a query compiler that generates optimal machine code for…
Rising cyber risks accelerated by generative artificial intelligence are making targeted digital disruptions increasingly probable, even though the global network's decentralized routing prevents an outright total collapse. Shared software dependencies and common sign-in services mean single vulnerabilities can still abruptly disable critical platforms like healthcare claims or banking applications. To prevent operational catastrophes, organizations must prioritize proactive incident response plans and test their…
The cloud hosting provider published an advisory regarding a newly tracked issue identified as CVE-2026-98432. To ensure affected customers stay informed as details emerge, the platform established dedicated notification channels via email and SMS. This outreach effort enables system administrators worldwide to monitor remediation efforts directly.
The technology giant is currently evaluating potential involvement to provide digital defense expertise for the Department of Defense. Participating in this defense initiative would expand the company's footprint in federal contracting, though it must balance government defense priorities with ongoing scrutiny from internal and external stakeholders.