===============Install ALACRITY=========================== DEPENDENCES: ===gcc 4.8.2 ===automake 1.11.1 autoconf 2.69, and libtool 2.6.2b You can download them from links below http://ftp.gnu.org/gnu/automake/ http://ftp.gnu.org/gnu/autoconf/ https://ftp.gnu.org/gnu/libtool/

1. Build timer package [git@github.com:ncsu-samatova/Timer.git] 1.1 cd [TIMER SOURCE CODE] 1.2 ./configure --prefix=[TIMER INSTALL PATH] 1.3 make ; make install
2. Build PForDelta package [git@github.com:ncsu-samatova/PForDelta.git] 1.1 cd [PForDelta PATH] 1.2 make libridcompress.a // this will produce libridcompress.a in current folder
3. Build ALACRITY [git@github.com:ncsu-samatova/ALACRITY.git] 1.1 cd [ALACRITY PATH] 1.2 Copy ./runconf to ./myrunconf 1.3 Edit ./myrunconf, setting the paths for timer and indexcompression, and your output build path 1.4 Run ./myrunconf

==============Index & Query ============================

1. Build Index: 1.1 cd [alalrity install path] 1.2 The executable file 'alac' is the command we need. It has following arguments ( type ./build/bin/alac, we can see the detail usage): 'encode': indicates to indexing the data -p : partition size, we can specify total size or total element number -i | -x : the format of index, -i means the inverted index, but -x means the compressed inverted index -s[Number] : the number of bits it will use [input data path] [output data path] : this path require a prefix of the files the alarity will produce For example, if we want to build compressed inverted indexes for 2GB double temp precision data by binning 16 significant bits csc$ ./build/bin/alac encode -p268435456E -x -edouble -s16 ~/sigmod_2012_2GB/temp ~/data/ii_index/s3d/1part_2G/pfd/temp Using partition size of 268435456(not accounting for element size, but WILL later) Using legacy format n Encoding complete, 2147483648 bytes of input data successfully encoded into 1 partitions

2. Perform Query: 2.1 cd [alalrity install path] 2.2 The executable file 'uniquery' is the command we need. It has following arguments ( type ./build/bin/uniquery, we can see the detail usage): [input data base path]: the path is the output path of the indexing stage [low boundary value] [high boundary value] For example, if we want to query the temp data is between 35.65 and 663.6, we do: csc$ ./build/bin/uniquery ~/data/ii_index/s3d/1part_2G/pfd/temp 35.65 663.6 Total results: 11041898 (4.1134% selectivity, 1/1 partitions had results) uniquery avg. time 0.261354

===============ASSUMPTIONS=========================================

1. ALACRITY now only takes at most 2GB partition data size.
2. The output of ALACRITY query does not output all RIDs.