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

.

Browse Source
This commit is contained in:
powturbo
2017-06-13 10:03:48 +02:00
parent 384763a456
commit 1906c40643
1 changed files with 1 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@ -174,12 +174,10 @@ using [900.000 multicore servers](https://www.cloudyn.com/blog/10-facts-didnt-kn
- Recent "integer compression" GOV2 experiments (best paper at ECIR 2014) [On Inverted Index Compression for Search Engine Efficiency](http://www.dcs.gla.ac.uk/~craigm/publications/catena14compression.pdf) using 8-core Xeon PC are reporting 1.2 seconds per query (for 1.000 Top-k docids). - Recent "integer compression" GOV2 experiments (best paper at ECIR 2014) [On Inverted Index Compression for Search Engine Efficiency](http://www.dcs.gla.ac.uk/~craigm/publications/catena14compression.pdf) using 8-core Xeon PC are reporting 1.2 seconds per query (for 1.000 Top-k docids).
### Compile: ### Compile:
git clone --recursive git://github.com/powturbo/TurboPFor.git git clone --recursive git://github.com/powturbo/TurboPFor.git
cd TurboPFor cd TurboPFor
###### Linux, Windows (MingW), Clang,... ###### Linux, Windows (MingW), Clang,...
make make
or or
make AVX2=1 make AVX2=1
@ -191,10 +189,8 @@ using [900.000 multicore servers](https://www.cloudyn.com/blog/10-facts-didnt-kn
make NSIMD=1 make NSIMD=1
###### Windows visual c++ ###### Windows visual c++
nmake /f makefile.vs nmake /f makefile.vs
### Testing: ### Testing:
##### - Synthetic data (use ZIPF parameter): ##### - Synthetic data (use ZIPF parameter):
+ benchmark groups of "integer compression" functions <br /> + benchmark groups of "integer compression" functions <br />