Recap from last time

1
Recap from last time

• We saw various different issues related to
  program analysis and program transformations
• You were not expected to know all of these things
• We will learn more about each one of the topics
  we touched upon last time

2
Project schedule

• Find groups a week from today
• groups of 2-3
• groups of 1 possible if you ask me
• Project proposal due the following Tuesday
• but... how are we supposed to come up with
  ideas!!!!

3
Ideas for generating project ideas

• If you are currently doing research
• that is related to programming languages/compilers
  
• carve out some part of your current research, and
  do it for the project
• Think about what annoys you the most
• in programming
• and fix it using program analysis techniques

4
Ideas for generating project ideas

• Think about technology trends
• iphone, mobile devices
• multicore, parallelism
• web browsers, java scripts, etc.
• bluetooth, ad-hoc networks
• how do these change the way we program/compile?
• Try to find bugs in real programs using PL
  techniques
• search for Dawson Engler and look at some of
  his papers. Try to use similar techniques to find
  real bugs.

5
Ideas for generating project ideas

• How can I use information that is out there?
• ask yourself what information is out there? And
  how can I record it/analyze it/use it to improve
  programmer productivity?
• For example cvs logs, bugzilla, keystrokes,
  mouse movements, etc.
• Talk to your neighbor
• in the 231 class, that is
• have a discussion

6
Tour of common optimizations
7
Simple example
foo(z) x 3 6 y x 5
return z y
8
Simple example
foo(z) x 3 6 y x 5
return z y
9
Another example
x a b ... y a b
10
Another example
x a b ... y a b
11
Another example
if (...) x a b ...
y a b
12
Another example
if (...) x a b ...
y a b
13
Another example
x y ... z z x
14
Another example
x y ... z z x
15
Another example
x y ... z z y
What if we run CSE now?
16
Another example
x y ... z z y
What if we run CSE now?
17
Another example
x yz ... x ...
18
Another example
x yz ... x ...

• Often used as a clean-up pass

Copy prop
DAE
x y z z x
x y z z y
x y z z y
19
Another example
if (false) ...
20
Another example
if (false) ...
21
Another example

• In Java

a new int 10 for (index 0 index lt 10
index ) aindex 100
22
Another example

• In lowered Java

a new int 10 for (index 0 index lt 10
index ) if (index lt 0 index gt
a.length()) throw OutOfBoundsException
aindex 0
23
Another example

• In lowered Java

a new int 10 for (index 0 index lt 10
index ) if (index lt 0 index gt
a.length()) throw OutOfBoundsException
aindex 0
24
Another example
p x p 5 y x 1
25
Another example
p x p 5 y x 1
x 5 p 3 y x 1
???
26
Another example
for j 1 to N for i 1 to M ai
ai bj
27
Another example
for j 1 to N for i 1 to M ai
ai bj
28
Another example
area(h,w) return h w h ... w 4 a
area(h,w)
29
Another example
area(h,w) return h w h ... w 4 a
area(h,w)
30
Optimization themes

• Dont compute if you dont have to
• unused assignment elimination
• Compute at compile-time if possible
• constant folding, loop unrolling, inlining
• Compute it as few times as possible
• CSE, PRE, PDE, loop invariant code motion
• Compute it as cheaply as possible
• strength reduction
• Enable other optimizations
• constant and copy prop, pointer analysis
• Compute it with as little code space as possible
• unreachable code elimination

31
Dataflow analysis
32
Dataflow analysis what is it?

• A common framework for expressing algorithms that
  compute information about a program
• Why is such a framework useful?

33
Dataflow analysis what is it?

• A common framework for expressing algorithms that
  compute information about a program
• Why is such a framework useful?
• Provides a common language, which makes it easier
  to
• communicate your analysis to others
• compare analyses
• adapt techniques from one analysis to another
• reuse implementations (eg dataflow analysis
  frameworks)

34
Control Flow Graphs

• For now, we will use a Control Flow Graph
  representation of programs
• each statement becomes a node
• edges between nodes represent control flow
• Later we will see other program representations
• variations on the CFG (eg CFG with basic blocks)
• other graph based representations

35
Example CFG
x ...
y ...
x ... y ... y ... p ... if (...)
... x ... x ... ... y ... else
... x ... x ... p ... ... x
... ... y ... y ...
y ...
p ...
if (...)
... x ...
... x ...
x ...
x ...
... y ...
p ...
... x ...
... x ...
y ...
36
An example DFA reaching definitions

• For each use of a variable, determine what
  assignments could have set the value being read
  from the variable
• Information useful for
• performing constant and copy prop
• detecting references to undefined variables
• presenting def/use chains to the programmer
• building other representations, like the DFG
• Lets try this out on an example

37
x ...
Visual sugar
y ...
1 x ... 2 y ... 3 y ... 4 p ...
y ...
p ...
if (...)
... x ... 5 x ... ... y ...
... x ... 6 x ... 7 p ...
... x ...
... x ...
x ...
x ...
... y ...
p ...
... x ... ... y ... 8 y ...
... x ...
... x ...
y ...
38
1 x ... 2 y ... 3 y ... 4 p ...
... x ... 5 x ... ... y ...
... x ... 6 x ... 7 p ...
... x ... ... y ... 8 y ...
39
1 x ... 2 y ... 3 y ... 4 p ...
... x ... 5 x ... ... y ...
... x ... 6 x ... 7 p ...
... x ... ... y ... 8 y ...
40
Safety

• Recall intended use of this info
• performing constant and copy prop
• detecting references to undefined variables
• presenting def/use chains to the programmer
• building other representations, like the DFG
• Safety
• can have more bindings than the true answer,
  but cant miss any

41
Reaching definitions generalized

• DFA framework is geared towards computing
  information at each program point (edge) in the
  CFG
• So generalize the reaching definitions problem by
  stating what should be computed at each program
  point
• For each program point in the CFG, compute the
  set of definitions (statements) that may reach
  that point
• Notion of safety remains the same

42
Reaching definitions generalized

• Computed information at a program point is a set
  of var ! stmt bindings
• eg x ! s1, x ! s2, y ! s3
• How do we get the previous info we wanted?
• if a var x is used in a stmt whose incoming info
  is in, then

43
Reaching definitions generalized

• Computed information at a program point is a set
  of var ! stmt bindings
• eg x ! s1, x ! s2, y ! s3
• How do we get the previous info we wanted?
• if a var x is used in a stmt whose incoming info
  is in, then s (x ! s) 2 in
• This is a common pattern
• generalize the problem to define what information
  should be computed at each program point
• use the computed information at the program
  points to get the original info we wanted

44
1 x ... 2 y ... 3 y ... 4 p ...
... x ... 5 x ... ... y ...
... x ... 6 x ... 7 p ...
... x ... ... y ... 8 y ...
45
1 x ... 2 y ... 3 y ... 4 p ...
... x ... 5 x ... ... y ...
... x ... 6 x ... 7 p ...
... x ... ... y ... 8 y ...
46
Using constraints to formalize DFA

• Now that weve gone through some examples, lets
  try to precisely express the algorithms for
  computing dataflow information
• Well model DFA as solving a system of
  constraints
• Each node in the CFG will impose constraints
  relating information at predecessor and successor
  points
• Solution to constraints is result of analysis