lucy-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Apache Wiki <wikidi...@apache.org>
Subject [lucy-commits] [Lucy Wiki] Update of "LucyBookClub" by MarvinHumphrey
Date Wed, 09 Jan 2013 23:56:14 GMT
Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Lucy Wiki" for change notification.

The "LucyBookClub" page has been changed by MarvinHumphrey:
http://wiki.apache.org/lucy/LucyBookClub?action=diff&rev1=32&rev2=33

Comment:
Add Jan 15 2003 meeting on PFOR-DELTA paper.

  
  == Upcoming Meetings ==
  
- '''NOTICE: Lucy book club will not be meeting during the holiday season, next meeting is
January 3rd, 2013'''
- 
  === Tuesday January 15, 7:00 pm PDT: ===
+ 
+ The Lucy Book Club is taking a break from our book-in-progress this week to read a paper
on integer compression techniques.  One of the algorithms described in the paper is PFOR-DELTA
(Patched Frame-Of-Reference with delta encoding), which is particularly suitable for inverted
lists.
+ 
+ [[http://oai.cwi.nl/oai/asset/15564/15564B.pdf|Super-Scalar RAM-CPU Cache Compression]]
by Marcin Zukowski, Sándor Héman, Niels Nes, Peter Boncz
+ 
+ We'll go over the following questions:
+ 
+  1.  Why is PFOR-DELTA interesting to Lucy?
+  2.  What is a ''segment'' in PFOR-DELTA?
+  3.  What are the 4 major parts of a segment?  What is in them?
+  4.  What are the tradeoffs for choosing different sizes of `b` (bit width)?
+  5.  What type of data structure is used to keep track of exceptions?
+  6.  What is a ''compulsory exception''?  How does it influence your choice of `b`?
+  7.  PFOR-DELTA is a very interesting compression technique, but why is it really faster?
 What is PFOR-DELTA really optimizing for?
+  8.  In their testing, was RAM-RAM or RAM-Cache faster and why?
+  9.  Fine-grained access has some extra cost -- what is it?
  
  == Community Notes ==
  

Mime
View raw message