Live-variable analysis: Difference between revisions

Revision as of 11:43, 14 April 2014

Revision as of 14:43, 9 May 2014

Line 5:

It is a "backwards way" analysis. The analysis is done in a backwards order, and the dataflow [[confluence operator]] is [[set union]].

{| class="floatright"

<table>

|<trcode>

<td>

The set of live variables at line L3 is {<code>b</code>, <code>c</code>} because both are used in the addition, and thereby the call to <code>f</code> and assignment to <code>a</code>. But the set of live variables at line L1 is

only {<code>b</code>} since variable <code>c</code> is updated in L2. The value of variable <code>a</code> is never used. Note that <code>f</code> may be stateful, so the never-live assignment to <code>a</code> can be eliminated, but there is insufficient information to rule on the entirety of <code>L3</code>.

</td>

<td>

<code>

L1: b := 3;

L2: c := 5;

Line 19 ⟶ 12:

goto L1;

</code>

|}

The set of live variables at line L3 is {<code>b</code>, <code>c</code>} because both are used in the addition, and thereby the call to <code>f</code> and assignment to <code>a</code>. But the set of live variables at line L1 is

only {<code>b</code>} since variable <code>c</code> is updated in L2. The value of variable <code>a</code> is never used. Note that <code>f</code> may be stateful, so the never-live assignment to <code>a</code> can be eliminated, but there is insufficient information to rule on the entirety of <code>L3</code>.

</td>

</tr>

</table>

The dataflow equations used for a given basic block ''s'' and exiting block ''f'' in live variable analysis are the following: