The Curious Engineer

All blogs / Databases

Spiral

spiraldb.com

Turn raw robot data into exact, tested, and versioned datasets for training and evaluation. Search synchronized streams down to the moment and stream reviewed windows directly to GPUs.

DatabasesStorage / filesystemsPerformance / optimizationRust

’24’25’26
Posts
24 from 19 Jun 2024 to 21 Oct 2026
Posting
monthly (11 in the last year)
Kind
company blog
Feed
https://spiraldb.com/rss/feed.xml

2026 11

  1. Faster DuckDB Queries with Vortex
  2. Statistics are aggregates
  3. Fusing GPU Kernels with Dynamic Dispatch
  4. From S3 to GPU in One Copy
  5. Vortex: One Format for Any Shape
  6. The Storage Layer for the Composable Data Stack
  7. How Spiral avoids a 9× video pipeline slowdown
  8. Reversibility is the bottleneck
  9. Compile or Prefilter? Matching LIKE over Compressed Strings
  10. Pattern Matching on Compressed Strings without Decompression
  11. Have your cake and decompress it too

2025 10

  1. Announcing Spiral
  2. Vortex: a Linux Foundation Project
  3. Vortex on Ice
  4. Towards Vortex 1.0
  5. So you want to use Object Storage
  6. ALP Rust is faster than C++
  7. What if we just didn’t decompress it?
  8. Zone maps, or “queries go brrr”
  9. Data Layouts: Where Bytes Find Their Forever Home
  10. Logical vs Physical data types

2024 3

  1. Compressing strings with FSST
  2. Trick-or-treating with Vortex
  3. Life in the fastlanes