TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: TurboPFor – Bringing Fastest Integer Compression Incl. SIMD to Java

14 pointsby powturboalmost 10 years ago

1 comment

powturboalmost 10 years ago
- Direct Access w&#x2F;o decompression<p>- Fastest Variable Byte implementation<p>- Novel Variable Simple faster than simple16, better than simple8-b<p>- Scalar Bit Packing decoding as fast as SIMD-Packing<p>- Bit Packing incl. Direct Access&#x2F;Update w&#x2F; zero decompression<p>- Fastest and most efficient SIMD Bit Packing<p>- Fastest SIMD-Elias Fano implementation<p>- Novel TurboPFor (PFor&#x2F;PForDelta) with direct access or bulk decoding. More efficient than ANY other &quot;integer compression&quot; scheme.<p>- Java Critical Native Interface. Access TurboPFor incl. SIMD from Java as fast as calling from C.<p>-----------------------------------------------------------<p>* Inverted Index + Intersections<p>- Novel Intersections w&#x2F; skip intervals, decompress the min. #blocks<p>- 2000! queries &#x2F;sec on GOV2 (25 MB docid) on a SINGLE core<p>- Parallel Query Processing on Multicores. 7000! queries&#x2F;sec, quad core CPU