Lecture 1: 2015-01-05

Notes by: Brandon Woo, Eric Kuecks, and Chris Konstad

TA Information

Name Email Office Hours Room
Muhammad Taqi Mehdi TBD Boelter 2432
Diyu Zhou TBD Boelter 2432

Example Problem

$ ls -l big
-rw-rw-r-- 1 eggert faculty 9223372036854775000 Oct 6 11:21 big
$ grep x big
$
$ time grep x big
real 0m0.009s

1019 bytes/10-2 seconds = 1021 bytes/s = 8 x 1021 bits/s = 8 Zetabits/s

XKCD Comic

US freight industry carrying MicroSD = 0.5 Zetabits/s

grep is able to quickly look through the huge file because it makes use of lseek with SEEK_DATA. This works on a "holey" file (a file that seems to be very large, but contains huge regions of zeroes so does not store as much information as it claims). These files do not use as much memory as they appear to, which is why this enormous file is able to fit into memory a all. Using lseek, grep is able to bypass these large regions of zeroes that cannot possibly contain the target character and only search a small part of the file.

$ echo x >> big
$ grep x big
Binary file 'big' matches

Course Organization

OS News

Definitions

Common Systems Problems

  1. Incomensurate scaling
  2. Emergent properties
  3. Propagation of effects
  4. Tradeoffs
  5. COMPLEXITY

Homework

Read Chapters 1-2.3 for next lecture