# Computers and calculating pi:  A zillion-bit word?

**URL:** <https://boards.straightdope.com/t/computers-and-calculating-pi-a-zillion-bit-word/403049>\
**Category:** Cecil's Columns/Staff Reports\
**Created:** [May 7, 2007, 1:32am UTC](https://boards.straightdope.com/t/computers-and-calculating-pi-a-zillion-bit-word/403049 "2007-05-07T01:32:06Z")\
**Posts on this page:** 1\
**Page:** 2

<div class="post-metadata">

**Author:** ![groman](https://avatars.discourse-cdn.com/v4/letter/g/73ab20/32.png) [@groman](https://boards.straightdope.com/u/groman)\
**Post date:** [May 18, 2007, 4:55pm UTC](https://boards.straightdope.com/t/computers-and-calculating-pi-a-zillion-bit-word/403049/21 "2007-05-18T16:55:55Z")

</div>

[QUOTE=Bytegeist]  
You’d see the opposite happen: the variable **i** could be optimized away (it does nothing!), but not the function call. In fact, it would be illegal for the compiler to drop it in this case. The function “fork” _might_ have side effects, for all it knows. You couldn’t drop it without changing the meaning of the program.

[/QUOTE]

Of course you’re right about “i”. However, I was making a small joke at the expense of metacomputation theories, and referring to a supercompiling optimizer.

The joke being that a supercompiler that would (incorrectly) convert

```auto

while(1) { fork(); }

```

into

```auto

while(1) { }

```

could be fooled by an addition of arithmetic that did nothing.

_looks up at all the glares_

Fine, fine, I’ll go back to my cave.

[Previous page](https://boards.straightdope.com/t/computers-and-calculating-pi-a-zillion-bit-word/403049.md?page=1)