daiduo/TurboPFor-Integer-Compression
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Readme

Browse Source
This commit is contained in:
x
2017-02-26 14:22:11 +01:00
parent d31754916d
commit dc42151b0e
1 changed files with 7 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
README.md
View File

@ -31,7 +31,11 @@ TurboPFor: Fastest Integer Compression [![Build Status](https://travis-ci.org/po
- :sparkles: Fastest **"Elias Fano"** implementation w/ or w/o SIMD/AVX2
<p>
+ **Transform**
- :sparkles: Scalar & SIMD Transform: Delta, Zigzag, Transpose/Shuffle, Floating point<->Integer
- :sparkles: Scalar & SIMD Transform: Delta, Zigzag, Transpose/Shuffle
<p>
+ **Floating Point Compression**
- :new: (Differential) Finite Context Method FCM/DFCM floating point compression
- Using TurboPFor, more than 2 GB/s throughput
<p>
+ **Inverted Index ...do less, go fast!**
- Direct Access to compressed *frequency* and *position* data w/ zero decompression
@ -380,4 +384,5 @@ header files to use with documentation:<br />
- [Small Polygon Compression](https://arxiv.org/abs/1509.05505) + [Poster](http://abhinavjauhri.me/publications/dcc_poster_2016.pdf) + [code](https://github.com/ajauhri/bignum_compression)
- [Parallel Graph Analysis (Lecture 18)](http://www.cs.rpi.edu/~slotag/classes/FA16/) + [code](http://www.cs.rpi.edu/~slotag/classes/FA16/handson/lec18-comp2.cpp)
Last update: 04 FEB 2017
Last update: 26 FEB 2017